Introduction

In past decades, the deterioration of air quality caused by increased human activity and manufacturing has attracted wide concern worldwide1,2. This not only reduces the visibility of the air atmosphere3 but also significantly harms human health4 and endangers the sustainable development of society and the economy5. High concentrations of PM2.5 and PM10 would reduce atmospheric visibility and increase the occurrence of traffic accidents, while excessive exposure to polluted air could cause many kinds of cardiovascular and chronic respiratory diseases (e.g. asthma)6 or even lead to premature death and cancer2. High concentrations of SO2 and NOx could cause acid rain and bring serious adverse ecosystem effects, such as the corrosion of buildings, soil acidification, and damage to crops and the aquatic environment7,8. Therefore, it is urgent to reduce air pollutants and improve air quality.

Urban air quality is influenced by various factors, including socio-economic factors (e.g., the level of economic development, urban population, car ownership, fuel emission, and usage of fossil resources, etc.) and meteorological factors (e.g., temperature, relative air humidity, wind speed, and precipitation, etc.)9,10. Socio-economic factors are the main pollution sources affecting urban air quality11,12, while meteorological conditions also influence air quality when the main pollution sources are relatively stable13,14,15,16. For example, urban emissions from human activity and manufacturing could cause environmental problems, such as photochemical smog, ozone layer depletion, acid rain, toxic chemical pollution, and global climate warming17,18. Monthly or seasonal air quality variations are caused by pollutant-intensive emission sources and meteorological conditions13,19,20. Therefore, strict and effective regulations and measures for air pollution prevention have begun to be implemented worldwide21,22.

China has suffered serious environmental degradation and air pollution, accompanying rapid economic growth and urbanization in recent decades. Population growth, energy consumption, motor vehicle increment, and industrial dust emission have become China's main causes of air pollution23. Many studies have reported that air pollution strongly impacts people’s health and life24,25. In order to mitigate air pollution, many pollution reduction technologies and policies have been implemented to improve air quality. The control measures (e.g., renewable energy utilization, traffic control, and flue gas desulfurization and denitration) adopted by China's central and local governments have achieved remarkable results. However, air pollution is still very serious26,27, and has gradually become a hot topic of high concern in China28. Scientific evaluation of the temporal change and spatial differences in air quality characteristics can help the government to examine the air pollution status and dynamics, while systematic analysis of the factors affecting air quality could provide evidence for policymakers to formulate effective measures10, and to optimize the urban expansion and development pattern, and land use/land cover characteristics to improve air quality at city levels29,30,31,32.

The temporal characteristics of air quality and their driving factors have been analyzed across multiple spatial scales in China, mainly in megacities (e.g., Beijing, Shanghai, Guangzhou, and Shenzhen, etc.)23,33,34,35. However, the spatial heterogeneity in air pollutants has seldom been evaluated at city scales due to limited observation sites. As one of the important and core cities in China's Yangtze River Economic Belt, Wuhan's rapid industrialization and urbanization have achieved short-term gains at the expense of the environment. Assessing and exploring air quality in Wuhan will guide Hubei Province and even the Yangtze River Economic Belt to a certain extent. However, few studies have comprehensively examined the temporal characteristics of air quality in Wuhan city, especially the spatial variations characteristics.

In this study, the official data on daily Individual Air Quality Index (IAQI) of PM2.5, PM10, SO2, NO2, CO, and maximum 8-h average O3 (O3_8h), monitored at 22 stations throughout the mega-city of Wuhan from January 2016 to December 2020, were used to examine spatio-temporal characteristics of air pollution in Wuhan city. The main objectives of this study are: (1) evaluate the variations in the concentrations of PM2.5 (PM2.5_C), PM10 (PM10_C), SO2 (SO2 _C), NO2 (NO2 _C), CO (CO _C), and O3_8h (O3_8h_C) at monthly, seasonal, and yearly scales in the air in Wuhan city during 2016–2020; (2) examine the spatial differences in PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C at seasonal and yearly scales across different sites; and (3) analyze the influence of the main meteorological factors and socio-economic indicators on air pollutants in Wuhan city. The results should provide city-scale evidence on spatio-temporal characteristics of air pollutants and a scientific basis for taking effective measures and new policy proposals to improve air quality in Wuhan city.

Materials and methods

Study area

Wuhan has a subtropical monsoon, humid climate with four distinct seasons. The annual average precipitation is 1205 mm, and the annual average temperature is 15.8–17.5°C36. As one of central China's most important cities and mega-cities, Wuhan has a land area of 8569.15 km2. It is an important hub in China due to its many waterways, convenient transportation, and advantageous position. With the rapid rise in its scale, the central strategic fulcrum of Wuhan city has become increasingly significant. Rapid urbanization, high population growth, a large number of vehicles, and high energy consumption have all led to ecological degradation in recent decades in Wuhan37.

Sampling site and air quality datasets

Data on the daily IAQI of PM2.5, PM10, SO2, NO2, CO, and O3_8h from January 2016 to December 2020 at 22 monitoring stations were derived from the official website of the Wuhan ecological environment bureau, which is open-access (http://hbj.wuhan.gov.cn/). PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C were calculated from the IAQI. The locations of 22 monitoring stations of the air quality in Wuhan were showed in Fig. 1. The figure is generated using ArcGIS 10.4 (ArcGIS for windows, version 10.4, Environmental Systems Research Institute, Inc., USA). The DEM data was downloaded from Geospatial Data Cloud (https://www.gscloud.cn/sources/).

Figure 1
figure 1

Locations of the air quality monitoring stations in Wuhan city. Notes: 1, Hankou Huaqiao; 2, Hankou Jiangtan; 3, Hanyang Yuehu; 4, Jianghan Honglingjin; 5, South Jianghan; 6, Qiaokou Gutian; 7, Wuchang Ziyang; 8, Donghu Liyuan; 9, No. 182 Minzu Avenue; 10, Hongshan Dida; 11, Qingshan Ganghua; 12, Zhuankou district; 13, Caidian district; 14, Wujiashan; 15, Dongxihu district; 16, Huangpi district; 17, Xinzhou district; 18, Huashan Ecological Art Museum; 19, Hannan district; 20, Jiangxia district; 21, Huagong district; 22, Chenhu Qihao.

Full size image

The 22 stations include one control site (Chenhu Qihao), one industrial site (Huagong district), eleven urban central sites (Hankou Huaqiao, Hankou Jiangtan, Hanyang Yuehu, Jianghan Honglingjin, South Jianghan, Qiaokou Gutian, Wuchang Ziyang, Donghu Liyuan, No.182 Minzu Avenue, Hongshan Dida, Qingshan Ganghua), and nine suburban sites (Zhuankou district, Caidian district, Wujiashan, Dongxihu district, Huangpi district, Xinzhou district, Huashan Ecological Art Museum, Hannan district, and Jiangxia district). The control site was located in the national wetland reserve, representing the background level of air quality in Wuhan.

Meteorological and socio-economic datasets

Daily meteorological data of the Wuhan meteorological observation station (30°37' N, 114°08' E) were downloaded from the China Meteorological Data Sharing Service System (http://data.cma.cn/). The meteorological factors data mainly include air temperature (T, °C), relative air humidity (RH, %), precipitation (Prec, mm), and wind speed (w, m/s) from 2016 to 2020. The data on socio-economic factors were obtained from the Wuhan Statistical Yearbook (http://tjj.wuhan.gov.cn/). The socio-economic factors data mainly include the yearly gross domestic product (GDP, 100 million yuan), per capita regional GDP (PGDP, yuan/population), permanent resident population (PRP, ten thousand population), population density (PD, population/km2), road area (RA, × 104 m2), number of civilian vehicles (CV), per capita area of parks and green spaces (PPGA, m2), the green coverage rate of built-up areas (GRB, %), total energy combustion (EC, tons of standard coal/km2), coal combustion (CAC, tons of standard coal/km2), coke combustion (CKC, tons of standard coal/km2), crude oil combustion (COC, tons of standard coal/km2), fuel oil combustion (FOC, tons of standard coal/km2), and electric power combustion (EPC, kWh) from 2009 to 2020. The data from 2020 were excluded from the correlation analysis results due to COVID-19 in 2020.

Statistical analysis

Correlation analysis was used to analyze the relevance of the six pollutants and the influence of the main meteorological factors and socio-economic indicators on the six pollutants. Monthly and yearly data were mainly used to analyze the correlation between air quality, meteorological, and socio-economic factors. Before correlation analysis, independence, and normality tests were conducted on the data. Pearson correlation coefficients were calculated for the relationships between air pollutants and the meteorological and socio-economic data. The correlation between factors was regarded as statistically significant or highly significant when the P value was less than 0.05 or 0.01, respectively. Detailed methods can be referred to Chen et al.23. Data analysis was performed through SPSS 16.0 (SPSS for Windows, version 16.0, Chicago, Illinois, USA), and the spatial and temporal variations were plotted using SigmaPlot 10.0 (SigmaPlot for windows, version 10.0, San Jose, California, USA).

Results

Temporal variations in air pollutants

PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C showed similar monthly and seasonal trends. In contrast, O3_8h_C showed opposite monthly and seasonal trends (Fig. 2). At the monthly scale, PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C all showed a single-valley change pattern, with the minimum value in July and the maximum value in December or January. In comparison, O3_8h_C showed a double-peak change pattern, with the peak value in June and September (Fig. 2a). At the seasonal scale, the maximum values of PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C were in winter (December, January, and February) and the minimum values were in summer (June, July, and August), while the maximum value of O3_8h_C was in summer and the minimum value was in winter (Fig. 2b). At the yearly scale, the maximum values of the annual average PM2.5_C, PM10_C, SO2_C, and CO_C were in 2016; the maximum value of the annual average NO2_C was in 2017; while the maximum value of the annual average O3_8h_C was in 2019 (Fig. 2c). Annual average PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C were lower in 2020 compared to all the other years, which was mainly attributed to the strict lockdown in Wuhan in early 2020.

Figure 2
figure 2

Temporal variations in PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C at monthly (a), seasonal (b), and yearly (c) scales during 2016–2020.

Full size image

Spatial variations in air pollutants

The seasonal trends in PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C from 2016 to 2020 were similar at all 22 sites (Fig. 3), while the yearly trends in PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C from 2016 to 2020 showed small differences between the sites (Fig. 4). Overall, all the air pollutants showed distinct seasonal patterns at all sites (Fig. 3). PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C exhibited the lowest and highest concentrations in summer and winter, respectively, at all observation sites. In contrast to PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C, O3_8h_C exhibited the lowest concentrations in winter at all observation sites. During the observation year, PM2.5_C and PM10_C decreased year by year at each station. The PM2.5_C was highest in Qingshan Ganghua. The PM10_C was higher in Hankou Huaqiao, Jianghan Honglingjin, Qiaokou Gutian, Qingshan Ganghua, Wujiashan, Dongxihu district, and Huagong district than in other stations. The SO2_C, and NO2_C at several stations were highest in 2017 and lowest in 2020, while O3_8h_C was the highest in 2019 at some stations (Fig. 4). The NO2_C was lower while O3_8h_C and SO2_C were higher in suburban sites than that in urban central sites. The PM10_C, SO2_C, NO2_C, and CO_C were higher in suburban and urban central sites than in the control site (Fig. 4).

Figure 3
figure 3

Spatial differences in PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C at seasonal scale during 2016–2020 across different sites.

Full size image
Figure 4
figure 4

Spatial differences in PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C at the yearly scale during 2016–2020 across different sites.

Full size image

The correlation between air pollutants and other factors

Correlations between the six air pollutants

Significant correlations were found between the concentrations of the six air pollutants (p < 0.01), except for the relevance between O3_8h_C and PM10_C (p > 0.05) (Table 1). PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C showed significantly positive correlations with each other (p < 0.01), while O3_8h_C presented significantly negative correlations with PM2.5_C, SO2_C, NO2_C, and CO_C (p < 0.01).

Table 1 Correlations between the daily concentrations of the six pollutants during 2016–2019 (**p < 0.01; *p < 0.05).
Full size table

Correlations between air pollutants and influencing factors

The correlation between Temperature (T) and O3_8h_C was significantly positive (p < 0.01), while the correlations between the other pollutants and T were significantly negative (p < 0.01) (Table 2). Relative air humidity (RH) was negatively correlated with SO2_C and O3_8h_C (p < 0.05), while the correlations between the other pollutants and RH were insignificant (p > 0.05) (Table 2). Precipitation (Prec) was negatively correlated with all pollutants (p < 0.01) except O3_8h_C (Table 2). There was no significant correlation between air pollutants and wind speed (w) (p > 0.05) (Table 2).

Table 2 Correlations between the monthly concentrations of the six pollutants and meteorological factors during 2016–2019 (T temperature; RH relative air humidity; Prec precipitation; w wind speed; **p < 0.01; *p < 0.05).
Full size table

PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C presented a significantly negative relationship with GDP, PGDP, PRP, RA, CV, and GRB (p < 0.05 or p < 0.01). Significantly negative relationships were found between PM2.5_C, SO2_C, NO2_C, CO_C, and PD (p < 0.05 or p < 0.01). PM2.5_C displayed a significantly positive relationship with EC, CAC, CKC, and FOC. PM10_C displayed a significantly positive relationship with FOC. SO2_C displayed a significantly positive relationship with EC, CAC, CKC, COC, FOC, and EPC and NO2_C displayed a significantly positive relationship with CAC, CKC, and FOC. There was no correlation between CO_C and EC, CAC, CKC, COC, FOC, EPC, and between O3_8h_C and all the socio-economic indicators (Table 3).

Table 3 Correlations between the yearly concentrations of the six pollutants and socio-economic indicators during 2009–2019 (GDP gross domestic product; PGDP per capita regional GDP; PRP permanent resident population; PD population density; RA road area; CV number of civilian vehicles; PPGA per capita area of parks and green spaces; GRB green coverage rate of built-up areas; EC total energy combustion; CAC coal combustion; CKC coke combustion; COC crude oil combustion; FOC fuel oil combustion; EPC electric power combustion; **p < 0.01; *p < 0.05).
Full size table

Discussion

Temporal variations in air pollutants

The monthly and seasonal trends in PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C and O3_8h_C can reflect the formation processes of the air pollutants. According to the results, PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C all showed a single-valley change pattern, with the maximum value in winter and the lowest value in summer. The lower PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C in summer are mainly attributable to better wet removal and atmospheric boundary layer diffusion conditions (Fig. 5), and fewer combustion sources. In comparison, emissions from heat sources were important contributors to the higher PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C in winter38,39 because of the lowest temperature (Fig. 5). O3_8h_C showed a double-peak change pattern that mostly peaked in summer due to higher solar radiation and temperatures in summer time (Fig. 5) could contribute to the photochemistry activity and thus the formation of O340,41.

Figure 5
figure 5

Monthly air temperature and precipitation during 2016–2019 in Wuhan.

Full size image

The PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C decreased year by year and were lowest in 2020, mainly attributed to the effectiveness of the environmental protection and pollution control strategies23, and the strict lockdown in Wuhan in early 2020. The phenomenon of pollutant reduction during the COVID-19 lockdown in this study is consistent with that reported in previous studies42. The decrease in SO2_C in Wuhan city indicated an effective upgrading of key industrial sectors (electric power and steel, etc.), especially the ultra-low emission transformation of electric power, the elimination of small and medium-sized coal-fired boilers, the transformation of heating from coal to gas and electricity. This was consistent with the study by Li et al.26, which showed that the total SO2 emission in China in 2016 was 75% less than that in 2007. The decrease in CO_C and NO2_C was mainly attributed to the effective regulation of traffic-related and coal combustion emissions. These all prove the effectiveness of China's air pollution control policies in recent years43, especially the "Action Plan on Air Pollution Prevention and Control"44. A series of air quality improvement measures, such as the "Clear water and blue sky" project and "Strengthening Measures of Air Pollution Prevention and Control in Wuhan," have been implemented in Wuhan city in recent years. These air quality improvement measures focus on controlling automobile exhaust emissions and gradually raising road vehicle exhaust emission standards, accelerating the renovation of coal-fired boilers, and controlling pollution sources, such as dust and volatile organic compounds. However, the annual average O3_8h_C showed no clear decrease trend in this study, consistent with Chen et al.23 The temporal variations in air pollutants indicated that Wuhan's air pollution control measures have achieved certain results, but challenges still exist.

Spatial variations in air pollutants

The seasonal patterns of PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C and O3_8h_C were similar at all 22 sites, indicating that the formation processes of air pollutants were not significantly influenced by the locations. However, the pollutants showed small differences between the stations in the same year. The PM2.5_C and PM10_C were higher in urban and industrial sites and lower in the control site, indicating high particulate matter pollution at urban and industrial sites. Wang et al.42 also found higher PM2.5_C and PM10_C at industrial and urban sites than at mountainous sites. The SO2_C were higher in industrial sites, suggesting that industrial manufacturing processes were a significant pollution source of SO2 emission in Wuhan. The findings are consistent with Syafei et al.10. The NO2_C was lower, while O3_8h_C was higher in suburban sites. The finding that O3_8h_C was higher in suburban sites is consistent with Wang et al.42, while the lower NO2_C in suburban areas was mainly related to the lower traffic volumes there10. Previous studies have also shown that automobile exhaust is the primary source of urban nitrogen oxide pollution42. The annual average CO_C during 2016–2020 showed no spatial differences were not consistent with Syafei et al.10 and Wang et al.42. This might mainly be attributed to the consistency in the spatial distribution of pollutant emissions45 and the impact of land use/land coverage in Wuhan city. Therefore, the extent of development and land use/land coverage should be considered further to explore spatial differences in the concentrations of air pollutants. Thus, urban green space could be integrated, and landscape patterns could be optimized to improve air quality and decrease air pollutant concentrations.

Correlations between air pollutants and influencing indicators

Significant positive correlations between PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C suggested that they had originated from the same sources or were impacted by the same drivers23,46. This indicated that control measures could simultaneously decrease the concentrations of these pollutants23. However, the negative correlations between O3_8h_C and the other pollutants indicate the difficulty in controlling the six pollutants simultaneously46,47. Further studies should be explored to reveal the formation and control strategies of O3_8h in Wuhan.

Correlations between the six pollutants and the meteorological indicators suggested that the wet deposition process (e.g., scavenging and wash-out) attributed to the increase in Prec could reduce PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C48,49. The rise in air temperature, on the one hand, could intensify the activity of atmospheric molecules to some extent, leading to the diffusion of PM2.5, PM10, SO2, NO2, and CO. On the other hand, the rise in temperature could intensify the photochemical reaction in the atmosphere, leading to the rise of O3_8h_C. This also explains the monthly and seasonal pattern of PM2.5_C, PM10_C, SO2_C, NO2_C, CO_C, and O3_8h_C23,50. There was a significant negative correlation between RH and O3_8h_C, indicating that ozone was easy to accumulate under low humidity conditions, and ozone concentration decreased with increasing relative air humidity. Wind speed showed no significant relationship with the air pollutants in this study, showing that wind speed did not intensify air turbulence to improve air quality in Wuhan.

PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C had a significantly negative relationship with GDP, PGDP, PD, PRP, RA, CV, and GRB. PM2.5_C, SO2_C, and NO2_C displayed a significantly positive relationship with coal and energy consumption. The increase in PPGA and GRB and the decrease in EC, CAC, CKC, COC, FOC, and EPC have compensated for the adverse effects of the increase in GDP, PGDP, PRP, PD, RA, CV in the same period, leading to an overall decrease in PM2.5_C, PM10_C, SO2_C, NO2_C, and CO_C from 2009 to 2020 in Wuhan. An increase in PPGA and GRB mainly reveals the investments in environmental protection, while a decrease in EC, CAC, CKC, COC, FOC, and EPC mainly reveals the controls on anthropogenic sources, indicating that increase in atmospheric environmental carrying capacity and effective emission-cutting measures were effective in reducing the accumulation of PM2.5, PM10, SO2, NO2, and CO. Air pollutants were inversely proportional to vehicles ownership, which was attributed to the widespread use of clean energy vehicles. Compared to the other pollutants, the correlation between O3_8h_C and the socioeconomic factors suggested that the control of O3_8h_C is still a challenge for Wuhan. The urban sites with the highest population densities, high numbers of automobiles, and clustered commercial living zones have relatively stable air pollutant emissions. The air pollution caused by municipal emissions has become a dominant source in urban areas. The industrial site with high energy consumption intensity, leading to a high concentration of heavy pollution industrial activities. The air pollutants from urban and industrial sites were also transferred to the other sites, which was a major cause of Wuhan’s poor air quality.

Conclusion

This study examined the tempo-spatial characteristics of PM2.5_C, PM10_C, SO2 _C, NO2 _C, CO _C, and O3_8h_C, monitored at 22 stations in Wuhan city from January 2016 to December 2020, and their correlations with the meteorological and socio-economic factors. Based on 5-year data, we found that the maximum values of PM2.5_C, PM10_C, SO2 _C, NO2 _C, and CO _C were in winter, while O3_8h_C were in summer. The temporal variations in air pollutants showed that annual average PM2.5_C, PM10_C, SO2 _C, NO2 _C, and CO _C were lower in 2020 than in other years. The spatial variations in air pollutants indicated that the PM2.5_C and PM10_C were higher in urban and industrial sites and lower in the control site. The SO2_C was higher in industrial sites. The NO2_C were lower, and O3_8h_C was higher in suburban sites, while CO showed no spatial differences in their concentrations.

Significant correlations were found between the concentrations of the six air pollutants, except for the relevance between O3_8h_C and PM10_C. The correlations between O3_8h_C and the other pollutants were more complex than those between the other 5 pollutants. PM2.5_C, PM10_C, NO2 _C, and CO _C were significantly negatively correlated with temperature and precipitation but insignificantly associated with relative air humidity. SO2_C was significantly negatively associated with temperature, relative air humidity, and precipitation. O3_8h_C was significantly positively associated with temperature while significantly negatively associated with relative air humidity. There was no significant correlation between air pollutants and wind speed. Several socioeconomic factors drove air quality concentrations. The urban and industrial sites with high population densities, number of automobiles, and energy consumption intensity, led to a high concentration of pollutants. An increase in atmospheric environmental carrying capacity and effective emission-cutting measures effectively reduced the accumulation of PM2.5, PM10, SO2, NO2, and CO. Overall, this study provided scientific insights that tempo-spatial characteristics and major influencing factors should be taken into account to improve Wuhan’s air quality.