Temporal variations in ambient air quality indicators in Shanghai municipality, China

Official data on daily PM2.5, PM10, SO2, NO2, CO, and maximum 8-h average O3 (O3_8h) concentrations from January 2015 to December 2018 were evaluated and air pollution status and dynamics in Shanghai municipality were examined. Factors affecting air quality, including meteorological factors and socio-economic indicators, were analyzed. The main findings were that: (1) Overall air quality status in Shanghai municipality has improved and number of days meeting ‘Chinese ambient air quality standards’ (CAAQS) Grade II has increased. (2) The most frequent major pollutant in Shanghai municipality is O3 (which exceeded the standard on 110 days in 2015, 84 days in 2016, 126 days in 2017, 113 days in 2018), followed by PM2.5 (120days in 2015, 104 days in 2016, 67 days in 2017, 61 days in 2018) and NO2 (50 days in 2015, 67 days in 2016, 79 days in 2017, 63 days in 2018). (3) PM2.5 pollution in winter and O3 pollution in summer are the main air quality challenges in Shanghai municipality. (4) Statistical analysis suggested that PM2.5, PM10, SO2 and NO2 concentrations were significantly negatively associated with precipitation (Prec) and atmosphere temperature (T) (p < 0.05), while the O3 concentration was significantly positively associated with Prec and T (p < 0.05). Lower accumulation of PM, SO2, NO2, and CO and more serious O3 pollution were revealed during months with higher temperature and more precipitation in Shanghai. The correlation between the socio-economic factors and the air pollutants suggest that further rigorous measures are needed to control PM2.5 and that further studies are needed to identify O3 formation mechanisms and control strategies. The results provide scientific insights into meteorological factors and socio-economic indicators influencing air pollution in Shanghai.

With the growing need for improving air quality across cities, municipalities, and provinces in China, a series of laws, regulations, standards and control measures have been formulated and promulgated 1,2,4,8,23 . The ' Air Pollution Prevention Action Plan' was enacted on September 10, 2013, and the most stringent environmental protection law to date was implemented on January 1, 2015 8 . Significant measures have also been taken to mitigate the adverse effects of air pollution 24 . Air quality monitoring systems have been established in more than 330 cities 16 and at 1,300 national air quality monitoring sites 24 . Daily data on air quality index (AQI) and air quality indicators are released publicly on local government websites, providing an important foundation for air quality research and policy. In the past three decades, knowledge on air pollution has improved considerably with the growing number of publications on air pollution in megacities 2,4,8,14,16,22,24,25 . Many studies have reported spatio-temporal variations in particulate matter (PM 2.5 and PM 10 ) and gaseous (SO 2 , NO 2 , CO, and O 3 ) pollutants in Chinese cities 4,8,16,24,26 , and associated health and socioeconomic costs 3,6,14,21,22,[27][28][29] . Between 2013 and 2018, China's rigorous air pollution control greatly reduced the annual mean level of PM 2.5 in the atmosphere of 74 large cities 30 .
Shanghai is an important political, economic, and cultural center of China. With the acceleration of urbanization and industrial processes, Shanghai's environmental problems have become increasingly prominent, with air quality being one of the most serious issues. As a pioneer city in construction of ecological civilization, Shanghai's air quality has received much attention. In this study, official data on daily concentrations of PM 2.5 , PM 10 , SO 2 , NO 2 , CO, and maximum 8-h average concentration of O 3 (O 3 _8h) in the air in Shanghai municipality from January 2015 to December 2018 were used to examine air pollution status and dynamics in the municipality. The following aspects are addressed in this paper: (1) Temporal variations in average daily concentrations of PM 2.5 , PM 10 , SO 2 , NO 2 , CO, and O 3 _8h in the air in Shanghai municipality during 2015-2018; (2) annual and seasonal variations in major pollutants and number of days when concentrations exceeded the air quality standard; and (3) the main meteorological factors and socio-economic indicators affecting air pollution in Shanghai. The results were used to identify air quality management gaps in the municipality.

Results and discussion
Overview of air pollutants in Shanghai during 2015-2018. The average mass concentrations of the target pollutants during 2015-2018 were analyzed. We used the cumulative distribution of daily average values of PM 2.5 , PM 10 , NO 2 , SO 2 , CO, and O 3 _8h to determine the number of days during which Shanghai municipality was exposed to air pollution ( Fig. 1) 24 . For at least some half-days in 2015 (2016,2017,2018), Shanghai municipality was exposed to average values higher than 59 (50,45,40) μg m −3 for PM 2.5 , 52 (48,47,40) μg m −3 for PM 10 , 45 (43,47,44) μg m −3 for O 3 _8h, 48 (45,47,44) μg m −3 for NO 2 , 13 (12,9,8) μg m −3 for SO 2 , and 18 (18,18,15) mg m −3 for CO. This indicates a decrease in the number of days per year in which Shanghai residents were exposed to high concentrations of PM 2.5 , PM 10 , NO 2 , SO 2 , and CO. temporal variations in air pollutants. Following implementation of the six-round, 3-year environmental protection action plan, ambient air quality in Shanghai municipality has improved slightly. In 2018, the aver-   8,31 . Our results also indicated that more than 70% of the total mass of PM 10 was composed of PM 2.5 , which is close to the ratio reported in previous studies 8,24 . The decreases in CO and NO 2 concentrations were mainly attributable to effective regulation of coal combustion emissions and traffic-related emissions 8,31-33 . The reductions amplitudes were lower for CO and NO 2 compared with PM 2.5 , PM 10 , and SO 2 , which may be related to the rapid increase in vehicles in Chinese cities 8   The most frequent "major pollutant" in Shanghai municipality was O 3 , followed by PM 2.5 and then NO 2 and PM 10 . In comparison, SO 2 and CO were the "major pollutant" considerably less frequently. The number of days on which PM 2.5, O 3 , NO 2 , and PM 10 was designated the "major pollutant" was 120 (104, 67, 61), 110 (84, 126, 113), 50 (67, 79, 63) and 16 (13,13,14) in 2015 (2016, 2017, 2018), respectively. The low incidence of SO 2 as a "major pollutant" again indicated effective control of coal combustion and implementation of desulphurization systems 8,31 . Compared with 2015, the incidence of O 3 as a major pollutant in Shanghai increased to reach its highest value in 2017. This is consistent with the 90th percentile of O 3 _8h concentration, which also peaked in 2017. Previous studies have suggested that O 3 is a complex secondary pollutant related to solar radiation, NO x , volatile organic compounds (VOC), and vertical transport in the boundary layer 8 , factors that are difficult to control effectively 35,36 . While the number of polluted days with PM 2.5 concentrations over 75 μg m −3 decreased from 2015 to 2018, the complex mixture of PM 2.5 and O 3 in the air is still a challenge to continuous improvement of air quality in Shanghai municipality 8,24 .
There were seasonal variations in the concentrations of each pollutant (Fig. 4a), and thus the days on which the air quality standard was exceeded (non-attainment days) were not equally distributed throughout the year (Fig. 4b), which is consistent with findings in previous studies 24,37 . November, December, January, February, and March were the dominant months with non-attainment days for PM 2.5 in Shanghai municipality, while April, May, June, July, August, and September were the dominant months with non-attainment days for O 3 _8h. Overall, winter months had the largest number of polluted days and highest mean concentration of PM 2.5 , followed by spring, autumn, and summer, which is consistent with previous findings 16 . This trend has been mainly attributed to coal-fired heating of buildings 16,[38][39][40] . Summertime O 3 pollution in Shanghai was much more severe than in the other seasons (Fig. 4b), and the probability of O 3 _8h exceeding the CAAQS Grade II value was highest in July (11.25 ± 5.85 day), followed by August (6.25 ± 4.65 day), May (5.75 ± 3.2 day), and June (5.5 ± 1.29 day). This is consistent with findings in previous studies that summer is the O 3 episode season in Chinese megacity  Correlations between air pollutants. Different air pollutants were significantly correlated (p < 0.01) with each other, except for SO 2 and O 3 (Table 1). There were significant positive correlations between PM 2.5 , PM 10 , CO, SO 2 , and NO 2 , suggesting that these pollutants originated from the same sources (e.g., vehicle and coal emissions) or were impacted by the same drivers 24 . Therefore controlling traffic and coal combustion emissions might be a way of simultaneously decreasing the concentrations of these pollutants. O 3 was significantly positively correlated with PM, and negatively correlated with NO 2 and CO (p < 0.01). The correlation coefficients were weaker, however, which can mainly be attributed to the complex, nonlinear, and temperature-dependent chemistry of O 3 concentration 20,43 . This indicates difficulty in controlling O 3 concentration and merits further investigations on O 3 formation and control strategies in Shanghai municipality.

Correlations between air pollutants and meteorological factors.
Correlations between the six main pollutants and meteorological factors are shown in Table 2. The results suggested that temperature (T) significantly impacted accumulation of all six pollutants in Shanghai municipality, while precipitation (Prec) and relative air humidity (RH) may have affected accumulation of some pollutants. Of all the meteorological factors that significantly impacted pollutant concentrations, the correlations between meteorological factors and PM 2.5 , PM 10 , CO, SO 2 , and NO 2 were negative, while the correlations between meteorological factors and O 3 were positive. The concentrations of PM 2.5 , PM 10 , SO 2 , NO 2 , and CO displayed a significantly negative relationship with Prec (p < 0.05 or p < 0.01), suggesting that the wet deposition could mitigate air pollution by the scavenge and wash-out process 16,44,45 . Relative humidity was strongly positively correlated with Prec, leading consistently to significantly negative correlations between PM 10 , SO 2 and NO 2 and RH. The consistency in correlations between the pollutants and T, and that between the pollutants and Prec, was partly explained by the significantly positive correlation between Prec and T. This also explains why the average concentration of the pollutants PM 2.5 , PM 10 , SO 2 , and NO 2 during June-September was lower than in other months 46,47 . Wind speed (W) did not show any marked relationship with the air pollutants studied, indicating that W did not enhance air ventilation and turbulence and thus improve air quality.

Correlations between air pollutants and socio-economic indicators. Shanghai is undergoing
strong socioeconomic development, with the permanent resident population (PRP) increasing from 14.14 million in 1995 to 24.18 million in 2017, and the GDP of Shanghai municipality increasing from 251.8 billion RMB in 1995 to 3,063.2 billion RMB in 2017 34 (Fig. 5). In the same period, Shanghai municipality continuously increased its environmental protection and construction efforts, with rolling implementation of the six-round, 3-year environmental protection action plan. Green space area (GE) has increased, from 6,561 hm 2 in 1995 to 136,327 hm 2 in 2017, environmental investment (EI) has also increased, from 4.65 billion RMB in 1995 to 92.35 billion RMB in 2017, and total amount of smoke emissions (SE) and total exhaust sulfur dioxide emissions (SDE) has decreased from 207.8 thousand tons and 534.1 thousand tons, respectively, in 1995 to 47 thousand   (Fig. 5). Although ambient air quality in Shanghai municipality has improved slightly in recent decades as a result of its environmental regulations (Fig. 5), Shanghai is still one of the cities with the highest levels of air pollutants worldwide 48 .  Table 3. Although there have been large increases in PRP, GDP, EC, MV, and IEE in Shanghai in recent years, the increase in EI and the decrease in SE and SDE have compensated for the negative effects of the other factors, leading to positive effects in decreasing the concentrations of PM 10 , SO 2 , and NO 2 . The results revealed that investments in environmental protection and pollution control strategies were the main factors affecting accumulation of PM 10 , SO 2 , and NO 2 , indicating that such strategies are effective in reducing air pollution. The control in SE and SDE, and increase in EI and GS may be masking the increase in EC, MV, and IEE, leading to significant decrease in PM 10 , and slight decrease in NO 2 and SO 2 . The increased vehicle emissions and main energy would also help explain the relative stability NO 2 and SO 2 levels. As a pioneering city in the construction of ecological civilization, Shanghai has implemented several master plans to optimize GS in integration with an environmental sustainability agenda 49 . The implementation of ecological redline policy in Shanghai municipality could guarantee that GS be increased systematically or stabilized at this level 50 toward increasing the air quality. However, due to the lack in more detailed emission data per activity sector for all the pollutants, it is difficult to provide more concrete and quantitative evidence of the reasons that are driving the changes in the air quality, and explain if changes in air quality are really happening or if industrial sources are just getting better at not emitting the pollutants being monitored. Further studies are needed to reveal the percentage contribution of emission sources and atmospheric processes to the emissions of the pollutants.

conclusions
This study analyzed temporal variations in the concentrations of air pollutants (PM 2.5 , PM 10 , O 3 , SO 2 , NO 2 , and CO), the major pollutant on polluted days, and the number of non-attainment days in Shanghai municipality from January 2015 to December 2018. Based on 4-year data from the Shanghai Environmental Monitoring Center, the overall status of air quality in Shanghai has improved. The number of days that met CAAQS Grade II standards increased from 258 in 2015 to 296 in 2018.
We found that SO 2 was rarely the "major pollutant", indicating effective control of coal combustion and implementation of desulphurization system in Shanghai municipality. However, PM 2.5 pollution in wintertime and O 3 pollution in summertime are still major challenges to air quality improvement in Shanghai municipality. Our findings suggest that the most frequent major pollutant in Shanghai municipality is O 3  Statistical analysis suggested that different air pollutants were significantly correlated with each other, apart from SO 2 and O 3 . Significantly positive correlations between PM 2.5 , PM 10 , CO, SO 2 , and NO 2 were observed, suggesting that these pollutants may have originated from the same sources (e.g., vehicle and coal combustion emissions) or were impacted by the same drivers. The correlation results suggested that temperature (T) significantly impacted accumulation of all six pollutants in Shanghai municipality, while precipitation (Prec) and relative air humidity (RH) affected accumulation of some pollutants. Lower accumulation of PM, SO 2 , NO 2 , CO and more serious O 3 pollution in Shanghai were revealed in months with higher temperature and more precipitation. The correlation between the socio-economic factors and the air pollutants suggest that further rigorous measures are needed to control air pollution in the city. Investments in environmental protection and pollution control strategies were the main factors reducing accumulation of PM 10 , SO 2 , and NO 2 , indicating that these strategies are effective in reducing air pollution. Overall, this study provided scientific insights into impacts of meteorological factors and socio-economic indicators on air pollution in Shanghai. Table 3. Correlations between pollutants and socio-economic indicators based on yearly data for the period 1995-2017. GS: green space area; IEE: total industrial exhaust emissions; SE: total amount of smoke emissions; SDE: total amount of exhaust sulfur dioxide emissions; PRP: permanent resident population; GDP: gross domestic product; EC: energy combustion; MV: number of motor vehicles; EI: environmental investment. **p < 0.01; *p < 0.05.  Table 4.
where IAQI is individual air quality index and p is pollutant; and where IAQI p is individual air quality index of pollutant p, C p is concentration of pollutant p, BP Hi is high-value pollutant concentration limit when close to C p (in Table 4), BP LO is low-value pollutant concentration limit when close to C p (in Table 4), IAQI Hi is the individual air quality index corresponding to BP Hi , and IAQI LO is the individual air quality index corresponding to BP LO . Data on the real-time daily average concentrations of PM 2.5 , PM 10 , CO, NO 2 , and SO 2 and the maximum 8-h average concentration of O 3 at nine national air quality monitoring stations (Fig. 6)  CO is measured using the non-dispersive infrared absorption method 8,51 , PM 2.5 and PM 10 are measured using the micro-oscillating balance method and the β absorption method 8,51 , and SO 2 , NO 2 , and O 3 are measured by the fluorescence method, the chemiluminescence method, and the UV-spectrophotometry method, respectively 8,51 . Correlation analysis (using SPSS 16.0) was applied to determine the relevance of the six pollutants, meteorological factors, and socio-economic indicators. Independence and normality tests were performed before the correlation analysis. Pearson correlation analysis was performed when the data were normally distributed, otherwise Spearman correlation analysis was applied. AQI = max IAQI 1 , IAQI 2 , IAQI 3 , . . . , IAQI p IAQI p = IAQI Hi − IAQI Lo BP Hi − BP Lo C p − BP Lo + IAQI Lo Table 4. Individual air quality index (IAQI) and corresponding pollutant concentration limit 52 . a 1-h average concentration limits of SO 2 , NO 2 , and CO are only used in real-time reporting, and the 24-h average concentration limits of SO 2 , NO 2 , and CO are used in daily reporting. b When 1-h average concentration limit of SO 2 is higher than 800 μg m −3 , the individual air quality index of SO 2 is not reported and the reported individual air quality index of SO 2 is calculated by 24-h average concentration limits. c When 8-h average concentration limit of O 3 is higher than 800 μg m −3 , the individual air quality index of 8-h average concentration of SO 2 is not reported and the reported individual air quality index of SO 2 is calculated by 1-h average concentration limit.  0  0  0  0  0  0  0  0  0  0  0   50  50  150  40  100  50  2  5  160  100  35   100  150  500  80  200  150  4  10  200  160  75   150  475  650  180  700  250  14  35  300  215  115   200  800  800  280  1200  350  24  60  400  265