Spatiotemporal distribution of ground-level ozone in China at a city level

In recent years, ozone (O3) pollution in China has shown a worsening trend. Due to the vast territory of China, O3 pollution is a widespread and complex problem. It is vital to understand the current spatiotemporal distribution of O3 pollution in China. In this study, we collected hourly data on O3 concentrations in 338 cities from January 1, 2016, to February 28, 2019, to analyze O3 pollution in China from a spatiotemporal perspective. The spatial analysis showed that the O3 concentrations exceeded the limit in seven geographical regions of China to some extent, with more serious pollution in North, East, and Central China. The O3 concentrations in the eastern areas were usually higher than those in the western areas. The temporal analysis showed seasonal variations in O3 concentration, with the highest O3 concentration in the summer and the lowest in the winter. The weekend effect, which occurs in other countries (such as the USA), was found only in some cities in China. We also found that the highest O3 concentration usually occurred in the afternoon and the lowest was in the early morning. The comprehensive analysis in this paper could improve our understanding of the severity of O3 pollution in China.

Some scholars have launched investigations on the spatiotemporal pattern of O 3 . In Nanjing, a unimodal peak was observed with the highest O 3 levels occurring from 14:00 to 15:00, and the O 3 concentration reached its maximum and minimum levels in the summer and winter, respectively 3 . Wang et al. 26 studied the ground-level O 3 concentrations of 6 major Chinese cities located on both sides of the Heihe-Tengchong line, and they found that ground-level O 3 concentrations exhibited monthly variability, peaking in summer and reaching the lowest levels in winter. The diurnal cycle reached a minimum in the morning and peaked in the afternoon. Some research has found that the O 3 distribution pattern is also related to terrain features 9,27 .
As previously mentioned, most of the studies on O 3 spatiotemporal patterns are carried out with a short time scale and low spatial resolution and generally focus on a specific city or a limited spatial region. To the best of our knowledge, there has been a lack of research on the spatiotemporal pattern of O 3 in China using a higher spatial resolution and long time-series datasets. Recently, China established a large-scale ground real-time air quality monitoring network, which provides data we can use to conduct research on the spatiotemporal distribution pattern of O 3 pollution nationwide.
In brief, this research makes the following contributions. First, we obtained the O 3 concentration data of 338 cities across China for more than three years, covering 1-Jan-2016 to 28-Feb-2019. In terms of spatial perspective, we investigated the O 3 concentrations in seven major geographic regions and three major urban agglomerations to conduct a more in-depth analysis and discussion. In terms of temporal perspective, we studied the annual, seasonal, monthly, weekly, daily, and diurnal and nocturnal variations in the O 3 concentrations. Second, the reasons for different patterns in different regions were briefly analyzed. The research results from this large dataset can not only help us elaborate on the spatiotemporal distribution pattern of O 3 concentration in China with a better spatiotemporal resolution and increase public awareness of the current O 3 pollution situation in China but also assist the relevant departments in formulating more targeted O 3 pollution prevention and control policies to meet the NAAQS and even the AQG standards in the future.

Results and Discussion
The NAAQS and the WHO set concentration limits for the maximum daily 8-hour average (MDA8) O 3 concentration. Two levels of limits are specified in the NAAQS (Grade 1 and Grade 2), and three levels of limits are specified in the WHO standard (AQG, Interim target 1 and High level) (see Table 1). Figure 1 shows the spatial distributions of the O 3 concentrations in 338 cities in China in 2016-2018. The regions with the most O 3 pollution were mainly concentrated in North China and Central China, especially in the Beijing-Tianjin-Hebei region (BTH) region. In addition, the O 3 pollution in the Chengdu-Chongqing and the Pearl River Delta region (PRD) regions was significantly higher than that of their neighboring regions. O 3 pollution in China has shown a trend of outward expansion. As shown in Table 2, based on the statistical results of the 90th percentile of the maximum daily 8-hour average urban O 3 concentration, the top 10 cities with severe O 3 pollution are mainly located in North China, Central China and the East China. Fig. 2 displays the over-standard rate of the O 3 concentration in seven geographical regions in China. None of the cities in North China met the AQG or Grade 1 limit, and nearly 70% exceeded the Grade 2 limit. In East and Central China, nearly 40% of urban O 3 concentrations exceeded the Grade 2 limit. The O 3 pollution in other regions was not so prominent; however, there were still considerable gaps from the AQG standard. O 3 is a secondary pollutant, which is generally formed in the atmosphere through photochemical pathways of NOx and volatile organic compounds (VOCs) [28][29][30][31][32] . Most of the NO x and VOCs come from heavy industries, such as coal-fired power plants, the steel industry, and the cement industry. Some studies found that the local photochemical reaction process has made an important contribution to the formation of O 3 33 , including the consumption of NO 2 during the photochemical reaction process 23 , which has been observed in regions such as North China and Yangtze River Delta (YRD) region 34,35 .

Spatial distribution of o 3 in china.
Additionally, PM pollution control in these regions has achieved certain results, and the reduction in haze has led to increased visibility, which in turn, has promoted the process of photochemical reactions and promoted the formation of O 3 pollution. It is worth noting that some cities in western China, where industrial activities are infrequent, sometimes have high concentrations of O 3 . In these high-altitude regions, the increase in O 3 concentration may be related to the transport of O 3 from the stratosphere to the troposphere 36 . In addition, meteorological environments with a high ultraviolet intensity and low humidity are conducive to O 3 formation. In general, the formation of O 3 pollution is affected by many factors, including prerequisite pollutant concentrations and meteorological conditions.  and the ozone levels were roughly the same in 2017 and 2018. Figure 1 shows that the scope of heavy O 3 pollution has gradually expanded. This phenomenon is also depicted in Fig. 4. In 2016, more than 95% of cities failed to meet the Grade 1 standard, and nearly 20% failed to meet the Grade 2 standard. In 2017 and 2018, these values increased to 99% and 30%, respectively.

Seasonal variation in o 3 in china.
The distributions of O 3 in different seasons were heterogeneous, exhibiting significant seasonal variations. In general, O 3 pollution in the summer is significantly higher than that in winter (Fig. 5). Because the photochemical reaction process is affected by meteorological conditions such as light and temperature, the meteorological conditions in summer are more suitable for photochemical reactions. In contrast, the UV intensity in winter is low, and the photochemical reaction is not enough to form heavy O 3 pollution. Seasonality is also reflected in spatial variation. In the spring and summer, O 3 pollution is highest in North, East, and Central China. In autumn, O 3 pollution gradually shifts to the south. In winter, national O 3 pollution is relatively mild, and only a small part of South China suffers from O 3 pollution. Overall, the problem of O 3 pollution in the eastern areas is more serious than that in the western areas. The seasonality of O 3 concentration  www.nature.com/scientificreports www.nature.com/scientificreports/ changes in the BTH region and the YRD region is relatively high. However, the seasonality of O 3 concentration changes in the PRD region is not as obvious. In the BTH region and the YRD region, the maximum and minimum O 3 concentrations were observed in the summer and winter, respectively. In the PRD region, the maximum O 3 concentration was observed in autumn, and the minimum was observed in winter.
The formation of O 3 pollution varies based on factors such as the overall NO x and VOC emissions 37-39 , topography 40 , and atmospheric circulation in the region 31,41 . Evidence suggests that the high O 3 pollution in the BTH region may be related to the emissions of precursor pollutants and the transportation of VOCs in neighboring provinces 42,43 . In the YRD region, the high temperatures in summer and the lower humidity can easily induce O 3 pollution. The O 3 concentration in the PRD region throughout the year is close and at a high level because the temperature throughout the year is similar and the annual average temperature exceeds 20°C in this region. Figure     www.nature.com/scientificreports www.nature.com/scientificreports/ winter to a certain degree. However, the general weekend effect of O 3 pollution is not significant, from a national scale. The weekly variation in O 3 concentration is affected by complex factors, the most likely of which is characteristics of urban resident activities. At present, no natural process has been found to produce climate change with a cycle of about 7 days, so Dominique et al. 54 believe that the existence of such a cyclic process is manifestation of human impact on climate. Due to the obvious weekly cycle of human activities, many meteorological elements in many regions have corresponding weekly cycle characteristics 55 . Meteorological elements of some cities have been observed to have varying degrees of weekly cycle characteristics, such as temperature 56,57 , precipitation frequency 58,59 , etc., which have a significant cycle with 7-day. The change of these climate factors will further affect www.nature.com/scientificreports www.nature.com/scientificreports/ the generation of O 3 in the photochemical reaction process, and thus affect the weekly variation. In general, the weekly variations in O 3 concentration are not very prominent, which shows that the weekly changes in human activities have limited effects on O 3 concentration. Figure 9 shows the daily O 3 concentration from January 1, 2016, to December 31, 2018. As shown in the figure, the daily variation in the O 3 concentration is usually continuous. The change from high concentration to low concentration, or from low concentration to high concentration, is often a gradual process rather than a sudden change. In most parts of the country, the daily variation of O 3 concentration shows an inverse U-shaped trend during each year, i.e., gradually increasing first and then decreasing. Except for South China, including the Pearl River Delta, the daily variation process of O 3 concentration has volatility. When observing horizontally from three years, the three cycles of O 3 variation can be clearly distinguished. We also found that for at least 1/3 of the days in the three years in each region, the O 3 concentration exceeded the AQG, while for more than 1/3 of the days in North China, the O 3 concentration exceeded Grade 1 of the NAAQS. When observing vertically, during the days with O 3 pollution, the BTH, YRD, and PRD regions usually had even higher O 3 concentrations than their neighboring areas. In short, the figure simultaneously shows the seasonal variation pattern as well as the spatial distribution characteristic of O 3 concentration.   www.nature.com/scientificreports www.nature.com/scientificreports/ of the important issues related to O 3 generation mechanisms, which will be helpful in determining the control of targeted emissions to reduce O 3 pollution 61 and formulating O 3 pollution control strategies.

Diurnal and nocturnal variation in o 3 in china. The hourly data on O 3 concentration are shown in
In this paper, we summarize the O 3 -NO x -VOC sensitivity regimes in major cities in China that have been studied, and the results are shown in Supplementary Table S1. In the urban districts of most cities, including Beijing, Tianjin, Shanghai and Guangzhou, O 3 generation is VOC-sensitive, mainly because human intervention in urban districts has greatly affected the emissions of precursors. Industry and transportation caused a large amount of NO x emissions, and the titration effect suppressed the increase in the O 3 concentration in urban areas. In these areas, the priority control of VOC emissions is more helpful in controlling local O 3 pollution. However, in the suburban areas of some cities, such as Lanzhou, Guiyang, Chongqing, and Xuzhou, the generation of O 3 is NO x -sensitive. The suburbs are less affected by anthropogenic emissions, and the migration of pollutants caused by the wind will affect O 3 pollution in the suburbs. In these areas, to suppress O 3 generation more effectively, priority should be given to the control of NO x emissions.  www.nature.com/scientificreports www.nature.com/scientificreports/ In addition, the meteorological influencing factors in major cities in China are provided in Supplementary  Table S1. The main meteorological factors that affect O 3 generation include temperature, relative humidity, wind speed, wind direction, solar radiation, atmospheric pressure, cloud cover, sunshine duration, precipitation, ultraviolet radiation, visibility, and geopotential height. The statistics of their frequency are shown in Supplementary  Fig. S1. In different regions, meteorological factors have heterogeneous effects on O 3 generation. In general, O 3 has a significant correlation with temperature and relative humidity. High temperature and low relative humidity are more conducive to the formation of O 3 , while meteorological factors such as sunshine duration, wind direction and wind speed have a crucial impact on the changes in O 3 concentration.
Combined with the results of previous statistical analyses, we found that the O 3 pollution affecting other cities is often caused by the synergistic effects of precursors and meteorological factors. For example, MDA8 in Beijing and its surrounding areas mainly occur at conditions of high temperature, low cloud cover, low relative humidity, weak southeast wind, low planetary boundary layer height, and the presence of a large amount of NO x and VOCs 62 . In Taiyuan, when the wind direction is southerly or southwesterly, the concentration of O 3 is higher, which indicates that the increase in O 3 concentration in Taiyuan is not only related to the local generation but also related to the external transport from the south 63,64 . The O 3 volume fraction and its generation rate in Langfang showed a significant positive correlation with air temperature and a significant negative correlation with total cloud cover. It is also susceptible to transmission in the southern region of Hebei and Tianjin.
From a long-term perspective, according to the characteristics of different O 3 pollution in different regions, priority should be given to strengthen the coordinated control of the sensitive precursor emissions in the region. Forecasting in advance when meteorological conditions are adverse and taking timely NO x and VOC control measures are important ways to solve regional O 3 pollution problems.

conclusions
This study analyzed the spatiotemporal distributions of O 3 concentrations in 338 prefecture-level cities in China from January 2016 to February 2019. The purpose was to understand the current status of O 3 pollution in China with a higher spatial resolution and a longer time series. Our study has the following findings: O 3 had obvious spatial heterogeneity. Only a few cities met the AQG standard of the WHO. O 3 pollution in North, East, and Central China was more serious, especially in the BTH region. The O 3 concentrations in the BTH, YRD, and PRD regions were usually higher than those in their neighboring cities. In the spring and summer, O 3 pollution in the north was more serious; in autumn, O 3 pollution shifted toward the south. In winter, the O 3 pollution problem was relatively mild across China. O 3 showed a significant temporal variation pattern. The O 3 concentration increased each year from 2016 to 2018. For the monthly variation in O 3 , except for South and Southwest China, other regions showed an inverted-V curve. Although the weekly variation in O 3 concentration was not exactly the same in different areas, some cities showed a W-shape. The O 3 concentration was lower on Tuesday and Saturday, and no obvious weekend effect was found. The study also characterized the diurnal and nocturnal variation pattern of O 3 concentration. The O 3 concentration was significantly higher during the day because of factors such as solar radiation, temperature, and precursor emissions. Due to the different time zones in different cities, the western region had a remarkable lag effect compared with the eastern region.
At present, China has made some achievements in the control of PM, NOx and other pollutants; however, the problem of O 3 pollution has become increasingly prominent. Against the background of China's severe composite air pollution, the need for the coordinated control of multiple pollutants is becoming increasingly apparent. According to our understanding, there is coexistence of VOC control and NOx control in China's O 3 pollution, and the reduction of particulate matter pollution has exacerbated the problem of O 3 pollution in China. The government should strengthen the monitoring of VOCs and combine the characteristics of O 3 pollution in different regions to formulate more targeted O 3 pollution control strategies to achieve a win-win situation of haze governance and O 3 control.