Analysis of Ozone Pollution Characteristics and Transport Paths in Xi’an City

Abstract

Xi’an is a key city for air-pollution prevention and control in China, and its near-ground ozone (O₃) pollution has become a key issue whose resolution is urgent. The spatial and temporal variations in ozone pollution and their relationship with meteorological factors, transport pathways, and potential source distribution in Xi’an City were investigated in this study using the backward trajectory clustering analysis, potential source contribution function (PSCF), and concentration weight trajectory (CWT) methods coupled with the hourly ozone-mass concentration observations and meteorological data from 2014 to 2020. The results indicate that Xi’an City has suffered from increasingly severe ozone pollution in recent years. Overall, the annual average mass concentration of O₃_8h_max presented an upward trend, exhibiting a 34.94% increase from 2014 to 2020. A seasonal variation peak occurred in summer, and the monthly variations featured an inverted “V” shape. Furthermore, the diurnal variation was significantly affected by the near-surface atmospheric photochemical process, showing a discernible single-peak and single-valley distribution with a peak between 15:00 and 17:00. The O₃ concentration demonstrated a significant linear positive correlation with surface temperature and sunshine duration. When the relative humidity was 30~50%, the temperature was above 30 °C, and the wind speed was ≤4 m/s, high-concentration O₃ pollution was liable to occur. The transmission mainly relied on short-distance airflow during periods of heavy O₃-polluted weather. In this context, in addition to local O₃ generation, high O₃ concentrations were also affected by the large NO_x and VOC emissions from heavy industries in neighboring cities in Shaanxi Province, southwest Shanxi Province, and northern Henan Province. Therefore, joint prevention and control measures on the O₃ pollution in Xi’an City must be adopted throughout the Fenwei Plain area with the aim of strictly controlling the emissions from regional pollution sources.

1. Introduction

Near-ground ozone (O₃) pollution is an important source affecting the quality of the atmospheric environment. It is mainly generated by precursor pollutants (such as VOCs, NO_x, etc.) generated by anthropogenic and natural sources through a series of photochemical reactions under the condition of solar radiation [1,2]. High concentrations of O₃ harm the growth of vegetation and crops and seriously affect human health [3,4,5,6,7]. Urban air-quality environmental monitoring data show that in recent years, China’s near-ground O₃ concentration has shown a significant increasing trend [8]. Under the premise of further control of fine particulate matter, the days of O₃ as the primary pollutant have increased year by year. Therefore, O₃ pollution has become another important problem that needs to be solved urgently as part of China’s air pollution control.

The research results of previous studies have shown that the phenomenon of near-ground O₃ concentrations exceeding the standard value in China is relatively common, and urban and regional O₃ pollution is intensifying [9,10]. For example, Xu et al. [11], Yin et al. [12], and Ma et al. [13] found that the concentration of O₃ showed an upward trend in Shanghai, Guangzhou, and the northern suburbs of Beijing, respectively. Yi et al. [14] researched the change characteristics of the O₃ concentrations of 25 cities in the Yangtze River Delta in 2013, and the results indicated that the spatial distribution of the O₃ pollution showed obvious patchy distribution characteristics, and the pollution in Shanghai and its surrounding cities was the most serious. Hu et al. [15] studied the variation characteristics of the O₃ concentrations in the Beijing–Tianjin–Hebei region and concluded that the O₃ concentrations increased over the years from 2014 to 2018, and the concentrations’ growth rates in different regions were significantly different; the O₃ high-value belt areas were concentrated in coastal cities, such as Tianjin, Qinhuangdao, Tangshan, etc. With the increase in the proportion of cities with O₃ exceeding the standard value, an increasing number of studies have begun to focus on the relationship between the spatial and temporal distribution characteristics of O₃ pollution and meteorological factors and other environmental factors, such as NO_x [16,17,18]. The research on atmospheric O₃ concentration and meteorological factors shows that high O₃ pollution events are usually related to meteorological conditions, such as high temperatures, strong radiation, low wind speed, and low relative humidity [19,20]. Gu et al. [21] found that the O₃ concentration increased more rapidly with the increase in temperature when the temperature in Shanghai was higher than 30 ℃ in summer. Wang et al. [22] found that temperature had the greatest impact on the O₃ concentration in Beijing, followed by air pressure and humidity. Zhao et al. [23] found that sunny days with less rain, higher temperature, stronger radiation, lower relative humidity, and lower wind speed were typical meteorological characteristics of O₃ concentrations above the standard values in Hong Kong.

Near-ground ozone has a relatively long atmospheric lifetime and regional transport characteristics. [24]. The local O₃ concentration is not only affected by the photochemical reaction of locally emitted precursors, but also by the transport of externally generated O₃ or O₃ precursors. Scholars have used the hybrid single-particle Lagrangian integrated trajectory (HYSPLIT), potential source contribution function (PSCF), and concentration weight trajectory (CWT) methods to study the cross-regional transport of O₃. For example, Dimitriou et al. [25] used backward trajectory analysis to find that the potential O₃ transboundary sources for Athens were mainly localized over the Balkan Peninsula, Greece, and the Aegean Sea, while Ioannina was mainly influenced by northern Greece and the Balkans Peninsula, and the difference in the pollution transmission between the two cities was mainly influenced by topography. Sharma et al. [26] used the PSCF and CWT methods to find that the Ganges Plain was the major potential source area of O₃ pollution in Delhi, India. Zhang et al. [27] used the backward trajectory and PSCF methods to find that O₃ pollution in Hangzhou was affected by the transport of pollutants in Zhejiang, Jiangsu, and Anhui Provinces. Fu et al. [28] discussed the potential source area of an O₃ pollution process in Hainan Province, and the results showed that Guangdong Province was the major potential source area of O₃ pollution in Hainan Province.

To sum up, in China, the current studies about the variation characteristics of O₃ pollution, its relationship with meteorological factors and transport pathways, and its potential source are mostly concentrated in North China, East China, and coastal cities. Although, against the background of China’s vigorous implementation of the Western development strategy, the urban economy in the central and western regions has developed rapidly, and this has caused serious atmospheric environmental problems, such as O₃ pollution. Therefore, it is necessary to carry out research on the characteristics of O₃ pollution, its transport path, and the distribution of potential sources in central and western cities.

Xi’an is the economic and cultural center of Northwest China. Its geographical location is between the northern Shaanxi Plateau and the Qinling Mountains. The terrain of the basin makes the natural diffusion capacity of air pollutants in Xi’an relatively weak. Moreover, in the past ten years, the rapid urbanization and industrialization of Xi’an have led to the increasing consumption of energy. However, this has been accompanied by the emissions of a large number of pollutants, such as NO_x and VOCs, into the atmosphere, resulting in increasingly serious O₃ pollution problems. The impact of O₃ on air quality is next only to PM_2.5 and PM₁₀, and the concentration of O₃ has risen rapidly in recent years. From 2016 to 2019, the 90th-percentile concentrations of O₃ during the 8 h of highest concentrations in Xi’an were 162, 185, 180, and 166 μg/m³, respectively [29]. Although more efforts have been made to control air pollution in recent years, Xi’an still featured in the bottom 20 in the air-quality rankings of 168 key cities in 2021, and the prevention and control of air pollution in Xi’an still require improvement.

Taking Xi’an as an example, this study first analyzed the O₃ concentration in Xi’an from 2014 to 2020 from the temporal and spatial change characteristics of different scales and analyzed O₃ concentration’s relationship with meteorological factors, such as daytime hours, surface temperature, relative humidity, wind speed, and O₃ concentration. Second, using the backward trajectory clustering, PSCF, and CWT analysis methods to explore the main transport pathways of O₃ in Xi’an during periods of heavy pollution, the potential source areas that affect O₃ pollution in Xi’an were revealed. This study is expected to provide data support for the prevention and control of air pollution in Xi’an and the joint prevention and control of that in the Fenwei Plain.

2. Materials and Methods

2.1. Data Sources

The hourly O₃ mass concentration and nitrogen dioxide (NO₂) data in Xi’an were obtained from China’s National Ambient Air Quality Monitoring Network from 2014 to 2020. The average value of the O₃ concentration data represents the overall O₃ mass concentration in Xi’an. In the process of data collection, when the data are missing or abnormal, it is necessary to eliminate abnormal data and then fill in the missing data through linear interpolation. The annual mean value of the mass concentration of pollutant O₃_8h_max (moving average of the maximum 8-h concentration) refers to the arithmetic mean value of O₃_8h_max on all days in the study year, the quarterly mean value refers to the arithmetic mean value of O₃_8h_max on all days in the research season, the monthly mean value refers to the arithmetic mean value of O₃_8h_max on all days in the research month, and the daily mean value refers to the moving average value of the daily maximum 8-h concentrations. A single-day pollutant O₃_8h_max concentration exceeding the standard means that the concentration is higher than 160 μg/m³, and hourly pollutant O₃ concentration exceeding the standard means that the concentration is higher than 200 μg/m³. Meteorological data, such as temperature, relative humidity, daytime hours, and wind speed were obtained from the hourly monitoring data of the Xi’an Meteorological Station. The descriptive statistical summary of meteorological data and air quality variables show in

Table 1. Descriptive statistical summary of meteorological data and air quality variables.

	O3 (μg/m3)	Temperature (°C)	Relative Humidity (%)	Wind Speed (m/s)	Sunshine Duration (h)
Mean	47.22	15.31	71.61	1.14	5.39
Max	47.22	38.9	99	10.3	12.9
Mode	4	11.1	99	0	0
Min	0	−9.7	5	0	0
Kurtossis	3.76	−0.89	−0.84	11.07	−1.54
Skewness	1.74	−0.03	−0.41	2.34	0.03
Standard deviation	45.42	9.92	22.68	0.89	4.46

. The meteorological data used in the backward trajectory (HYSPLIT) model were the Global Data Assimilation System (GDAS) data provided by the US Center for Environmental Prediction (NCEP), with a spatial resolution of 1° × 1°.

2.2. Research Method

2.2.1. HYSPLIT Model

The HYSPLIT model is an integrated model that has the function of dealing with various input fields of meteorological elements, various physical processes, and different types of pollutant emission sources, and was developed by the National Oceanic and Atmospheric Center in the United States and the Australian Bureau of Meteorology [30]. This model is a mixture of the Euler and Lagrangian diffusion models. The Lagrangian method is used for advection and diffusion calculation, and the Euler method is used for concentration calculation. It is widely used to analyze and calculate the transport path, settlement mode, and diffusion trajectory of air pollutants. In this study, the transport trajectories of air masses were calculated and clustered using the HYSPLIT model so as to track the transport paths of airflows arriving in Xi’an (34.23° N, 108.93° E) based on the Meteoinfo software and the GDAS data provided by NCEP. The backward trajectory of the air mass arriving in Xi’an every hour (0:00–23:00) each day from 2016 to 2020 during the heavy O₃ pollution period was calculated. The trajectory was delayed for 48 h, and the simulation height was 500 m above ground level.

2.2.2. PSCF

The PSCF is a method for judging the possible orientation and location of emission sources based on the conditional probability function, and the possible location of the emission source is given by combining the air-mass trajectory and the feature values. The PSCF function is defined as the conditional probability that the corresponding element value exceeds the set threshold when the air mass passing through the ij grid (i and j represent longitude and latitude, respectively) reaches the observation point. The larger the PSCF value, the greater the contribution of the grid point to the pollutant mass concentration of the starting point [31]. The grid area corresponding with the high PSCF value is the potential source area of O₃ in Xi’an, and the trajectories passing through this area are the transport paths that affect the O₃ concentration. The grid of the study area is divided into 0.2° × 0.2°. In this study, the exceeding threshold was set as the O₃ secondary mass concentration limit of 160 μg/m³. When the O₃ mass concentration corresponding to the trajectory is higher than the set exceeding threshold, the trajectory is a pollution trajectory. PSCF_ij is the ratio of the number of polluted tracks passing through a grid (P_ij) to the total number of tracks in the grid (Q_ij). The formula is as follows:

$$PSC{F}_{ij}=\frac{P_{ij}}{Q_{ij}}$$

(1)

Since PSCF is a function based on conditional probability, the error of PSCF can increase as the distance between the starting point and the sampling point increases. When n_ij is small, there is large uncertainty. To reduce the error, the weight function WPSCF is usually applied to reduce the uncertainty of PSCF [32].

$$WPSCF=W_{ij}\times PSC{F}_{ij}$$

(2)

Among which,

$$W_{ij}\{\begin{array}{c}1.00 80<Q_{ij} \\ 0.70 20<Q_{ij}<80 \\ 0.42 10<Q_{ij}<20 \\ 0.05 O_{ij}<10 \end{array}$$

(3)

2.2.3. CWT

The input data and applicable grid resolution for the PSCF and CWT methods are the same. PSCF methods typically use concentration thresholds to assess potential sources of pollutants, meaning that some tracks may have the same PSCF value when pollutant concentrations are slightly or well above the standard, and therefore, it may not be possible to distinguish between moderate and heavy sources of pollution. However, the CWT method can calculate the weight concentration of each track. In the process of meshing, each grid point can be assigned a weight concentration by calculating the average pollutant concentration corresponding to all tracks passing through the grid, thus overcoming the limitations brought by the PSCF method. A larger CWT value indicates a higher regional transmission concentration value [33]. The formula is as follows:

$$C_{ij}=\frac{1}{{{\displaystyle \sum }}_{T=1}^N{\rho }_{ijT}}{\displaystyle \sum }_{T=1}^N{C}_T{\rho }_{ijT}$$

(4)

C_ij is the average weighted concentration of each grid (i, j), T represents the trajectory, N is the total number of tracks, C_T is the pollutant concentration (μg/m³) when the track T passes through a grid (i, j), and ${\rho }_{ijT}$ is the time that the trajectory T stays in a grid (i, j). The CWT method also uses the same weight function (W_ij) to reduce the uncertainty, thus obtaining the WCWT values. Therefore, the main source areas of external contribution to O₃ pollution in Xi’an are the grid areas with higher C_ij values, and the pollution trajectories passing through these grid areas are the main transport pathways that cause the increases in O₃ concentration.

3. Results and Discussion

3.1. Variation Characteristics of O₃ Mass Concentration from 2014 to 2020

From 2014 to 2020, the annual average mass concentration of O₃_8h_max in Xi’an shows a rising and fluctuating trend (Figure 1a). The annual average concentrations were 65.58, 74.92, 87, 93.5, 89.5, 88.33, and 88.5 μg/m³, among which the highest concentration of O₃_8h_max was in 2017, and an increase of 34.94% was observed in 2020 compared to 2014. In addition, the number of days exceeding the standard rate of O₃ pollution in Xi’an generally shows an upward trend; in particular, the number of days exceeding the standard of O₃ in 2017 reached 68, an increase of 30 days compared to 2016. In addition, the earliest and latest dates of the O₃ concentration exceeding the standard value have continued to advance. The earliest dates of the O₃ concentration exceeding the standard value were 3 April, 18 April, and 27 April, respectively, in 2018, 2019, and 2020. The latest dates of the O₃ concentration exceeding the standard value were 30 August 2018, 8 September 2019, and 8 September 2020, with concentrations of 168, 171, and 162 μg/m³, respectively. In view of the current state of O₃ pollution, Xi’an City has carried out special rectification for volatile organic matter pollution and motor vehicle pollution in recent years, including for industrial enterprises, auto repair painting, catering fumes, oil and gas storage and transportation, decorative coatings, high-pollution and high-emission vehicles, etc. In the last three years, the number of days on which O₃ pollution exceeded the standard rate decreased significantly compared to 2017; however, the state of O₃ pollution prevention and control is still not optimistic, and sudden heavy O₃ pollution events often occur. Therefore, it is necessary to strengthen the efforts to reduce O₃ pollution.

The monthly variation trend of the O₃_8h_max mass concentrations in Xi’an shows an obvious inverted “V” shape (Figure 1b). They gradually increased from January to the highest value in June to July, with a concentration range of 97–181 μg/m³, and then began to decline. In 2017, the O₃ mass concentration presented an abnormally high value (181 μg/m³ in July), which was related to the high temperature that year (the highest temperature range, from June to August, reached 35.72–39.45 °C) and less precipitation. From June to July, the solar radiation was strong, the temperature was high, the sunshine duration was long, and the photochemical reactions were strong, resulting in a high O₃ mass concentration [34,35]. In winter, the temperature was low, the sunshine duration was short, and the photochemical reaction was weak, meaning that the high concentrations of particulate matter increased the aerosol optical thickness and reduced the photochemical reaction rate of O₃ [36]. Therefore, the mass concentration of O₃ was at its lowest from December to January.

Figure 2 shows the diurnal cycle trend of the hourly average mass concentrations of the pollutants, NO₂ and O₃, in different seasons in Xi’an from 2014 to 2020. It can be seen in the figure that the daily variation trend of the O₃ mass concentration in different seasons presents a single-peak and single-valley distribution. Especially in summer, the maximum diurnal ranges of the O3 concentration were reached. This is also confirmed by the maximum standard deviation in summer in

Table 2. The standard deviation of daily variations of O₃ concentrations in the four seasons from 2014 to 2020 in Xi’an City.

	Spring		Summer		Autumn		Winter
	O3	NO2	O3	NO2	O3	NO2	O3	NO2
Standard deviation of the mean	7.10	3.31	10.50	2.31	3.94	2.18	3.38	2.23

. The lowest values of the O₃ concentration appeared at 06:00–08:00. The highest values of the O₃ concentration appeared at 15:00–17:00 (the time at which the O₃ concentration reached its highest values in summer was earlier due to the sun rising earlier and the temperature rising faster) and the O₃ concentration gradually decreased after reaching the peak. Higher temperatures and stronger solar radiation were the main reasons for the higher O₃ mass concentrations in the afternoons. In this study, the diurnal variation of the O₃ precursor, NO₂, exhibited a bimodal distribution. The first peak generally appeared at 8:00–10:00, which was caused by the morning peak Xi’an. After sunrise, the solar radiation increased, and NO underwent a photochemical reaction to generate NO₂. With the weakening of the solar radiation, the generation rate gradually slowed down, and the NO and O₃ underwent a titration reaction at night to generate NO₂, resulting in the second peak of NO₂, which generally appeared at 20:00–23:00. In addition, a study showed that vehicles have become an important source of NO₂ [37]. After 17:00, vehicles produce a large amount of exhaust gas during the rush hour, and the concentration of NO₂ rises again. At night, NO reacts with O₃ to generate NO₂, while the photolysis reaction between NO₂ and O₃ stagnates, which is an important reason for the occurrence of the peak value of NO₂ at night.

3.2. Relationship between O₃ Pollution and Meteorological Conditions

The production of high concentrations of O₃ pollution requires certain meteorological conditions. Meteorological factors play a critical role in the formation, deposition, transport, and dilution of O₃. Large-scale circulation also exerts an important impact on the changes of O₃. The authors analyzed the relationship between the monitoring data of the O₃ concentrations and the meteorological data from September 2020 to August 2021. Among these, the daily data were used to analyze the relationship between the sunshine duration and the ozone concentration, and the hourly data were used to analyze the relationships between the surface temperature, relative humidity, and wind speed and the O₃ concentration.

3.2.1. Sunshine Duration

The influence of photochemical reactions on the O₃ concentration is mainly reflected by sunshine duration. The conditions of photochemical reactions are closely related to sunshine. With the increase in the sunshine duration, the rate of photochemical reaction increases correspondingly, thus accelerating the generation of O₃. As shown in Figure 3, with the continuous increase in the sunshine duration, the concentration of pollutant O₃_8h_max increased correspondingly, and the over-standard rate of the O₃_8h_max concentration also shows an upward trend. When the sunshine duration exceeded 8 h, the over-standard rate of O₃_8h_max concentration had a significant upward trend. Especially when the sunshine duration was more than 10 h, the over-standard rate of O₃_8h_max concentration reached 60%, and the average concentration was 170.4 μg/m³. In addition, the sunshine duration and the O₃ concentration show a significant positive correlation (the correlation coefficient is 0.66, with a significant correlation at the 0.01 level).

3.2.2. Temperature

O₃ can undergo a photochemical reaction under the action of solar radiation, and the change in the solar radiation intensity is reflected by the temperature. As shown in Figure 4, the average mass concentration of O₃ and its over-standard rate show an increasing trend with the increase in temperature. When the temperature was lower than 25 ℃, the O₃ mass concentration did not exceed the standard value. However, when the temperature exceeded 25 ℃, the O₃ concentration value appears to have exceeded the standard value. Especially when the temperature exceeded 30°, the over-standard rate reached 42.66%, and the average mass concentration was 153.72 μg/m³. The fitting analysis of the O₃ concentration and surface temperature shows a significant positive correlation (the correlation coefficient is 0.68, which is significantly correlated at the 0.01 level).

3.2.3. Relative Humidity

Under certain conditions, solar radiation is affected by water vapor in the atmosphere, which can affect photochemical reactions and consume O₃ in the atmosphere. As shown in Figure 5, the average mass concentration of O₃ increased first and then decreased with the increase in relative humidity, and the over-standard rate of the O₃ concentration shows the same trend. When the relative humidity was lower than 30%, the over-standard rate of O₃ concentration was relatively low. When the relative humidity was between 30–40% and 40–50%, a high O₃ concentration occurred frequently, and the over-standard rate reached 13.03% and 14.19%, respectively. However, when the relative humidity was greater than 60%, the over-standard rate dropped sharply, which is related to the fact that O₃ near the surface tends to produce wet deposition under the condition of high relative humidity, leading to a decrease in its concentration.

3.2.4. Wind Speed

Wind speed has a significant influence on the strength of stability in the atmospheric boundary layer. Different wind speeds also determine the removal efficiency and transport efficiency of pollutant O₃, and its action mechanism is relatively complex. As shown in Figure 6, when the wind speed was less than 4 m/s, the over-standard rate of O₃ concentration shows an upward trend with the increase in the wind speed. Among them, when the wind speed was between 3 and 4 m/s, the over-standard rate of O₃ concentration reached the highest value, reaching 10.15%. When the wind speed was between 5 and 6 m/s, the average O₃ mass concentration and over-standard rate began to decrease. When the wind speed was greater than 6 m/s, O₃ concentration did not exceed the standard. When the wind speed was small, it was conducive to the photochemical reaction and agglomeration of O₃ and its precursors, and then produced high concentrations of O₃ pollution. When the wind speed gradually increased, O₃ and its precursors may have been transported across regions, thus forming regional pollution. When the wind speed was greater than 6 m/s, the higher wind speed could diffuse and dilute O₃, leading to a decrease in its concentration.

The relationships between the temperature, relative humidity, and wind speed and the average mass concentration of O₃ are shown in Figure 7. The meteorological conditions for the formation of high concentrations of O₃ are: temperature ≥ 30 °C, relative humidity 30% to 50%, and wind speed ≤ 4 m·s⁻¹. The higher the temperature, the stronger the solar radiation and the more frequent molecular collisions, which are beneficial to the photochemical reactions that generate O₃ [38]. The conditions of high temperature and low humidity are conducive to the rapid generation of O₃. High temperatures usually occur in sunny weather conditions, with low water vapor content and less cloud cover in the atmosphere; consequently, the solar radiation reaching the ground increases, the rate of the photochemical reactions of O₃ precursors in the atmosphere increases, and the average mass concentration of O₃ increases. Higher relative humidity is not conducive to the formation of O₃ because there is sufficient water vapor under high-humidity weather conditions, and the possibility of cloud formation and precipitation is high, thereby weakening ultraviolet radiation and photochemical reactions [39,40]. An et al., (2009) found that when the relative humidity is around 60%, the photochemical reaction intensity has a limited value. When the relative humidity is too high, the photochemical reaction intensity of the O₃ precursor NO_x and CO decreases with the increase in relative humidity [41].

According to the previous analysis, the main meteorological factors affecting the O₃ concentration are temperature, sunshine, relative humidity, and wind speed. Therefore, the O₃_8h_max concentration (Y) was set as the explained variable, and the daily maximum temperature (X₁), sunshine duration (X₂), daily average relative humidity (X₃), and daily maximum wind speed (X₄) were set as the explanatory variables to establish a regression model. According to the regression analysis results, the estimation equation of O₃ concentration can be obtained:

$$Y=101.26+3.123X_1+4.591X_2-0.861X_3-1.125X_4-2.204$$

(5)

According to the regression equation, the higher the daily maximum temperature, the longer the sunshine duration, the smaller the daily average relative humidity, and the smaller the daily maximum wind speed in Xi’an City, the more beneficial it is to increase the O₃ photochemical reaction rate, that is, the higher the concentration of O₃_8h_max, the more likely an exceeding O₃ phenomenon will occur.

3.3. O₃ Transport Pathways and Potential Source Analysis in Xi’an during Heavy Pollution

Due to the high concentration of O₃ during the heavy pollution period, the variation characteristics of the O₃ concentration and the transport path of the external sources were more obvious. This section reports on the selection of the O₃ trajectories during the heavy pollution period from 2014 to 2020 for clustering and potential source analysis. The heavy pollution periods were as follows: 17–20 June 2016 (the O₃_8h_max average concentration was 217 μg/m³); 25–28 June 2017 (the O₃_8h_max average concentration was 219.25 μg/m³); 1–4 July 2017 (the O₃_8h_max average concentration was 239.5 μg/m³); 9–15 July 2017 (the O₃_8h_max average concentration was 219.86 μg/m³); 1–5 August 2017 (the O₃_8h_max average concentration was 216 μg/m³); and 5–9 July 2020 (the O₃_8h_max average concentration was 213.6 μg/m³).

During the periods of heavy pollution, the backward trajectory cluster analysis shows that Xi’an was dominated by short-distance transport airflow (Figure 8). Trajectory 1 originated from the southwest direction and reached Xi’an through the southwest of Shaanxi Province, accounting for a relatively high proportion of 25.12%. Trajectory 2, which had a longer conveying distance, originated from the northwest direction and reached Xi’an via Ningxia, eastern Gansu, and central Shaanxi, and features the lowest proportion of 13.33%. Trajectory 3 originated in the vicinity of Xi’an City. It features the shortest conveying distance and the highest proportion of 45.36%. Trajectory 4 originated from the northeast direction and reached Xi’an via southern Shanxi, northern Henan, and eastern Shaanxi, accounting for 16.19%. During the periods of heavy O₃ pollution, the air mass in Xi’an mainly came from the surrounding areas and the northeast and southwest of Xi’an. These air masses had a short transmission distance and slow-moving speed, and easily carried O₃ precursor pollutants (NO_x and VOCs) on the path to Xi’an, which had a major impact on the O₃ concentration in Xi’an through their superimposition on local pollution emissions.

During the periods of heavy pollution, the analysis results of the potential sources of O₃ pollution in Xi’an show (Figure 9) that the areas with high WPSCF values (>0.5) were located in the southwest and southeast of Shaanxi Province, the northwest of Henan Province, the southwest of Shanxi Province, and the central region of Shaanxi Province. In addition, the WPSCF value in the surrounding areas of Xi’an was also relatively large (>0.4), and its influence cannot be ignored. The WCWT results (Figure 9) show that the areas with the greatest contribution to the O₃ concentrations in Xi’an during heavy pollution were mainly concentrated in the areas surrounding Xi’an; they were mainly distributed in the southwest and south of Shaanxi Province and the northeast area of Xi’an, with CWT values > 90 μg/m³. Xi’an City is located in the center of Guanzhong, between the northern Shaanxi Plateau and the Qinling Mountains, in the middle of the long and narrow section of the trumpet basin, with a trumpet-mouth-shaped topography to the northeast. A center of the coal chemical industry is located in the southwest of Shanxi Province, upwind of the east and northeast of Xi’an City. Consequently, large volumes of pollutants, such as NO_x and VOCs, are discharged, and the polluted air flow easily reaches Xi’an along the trumpet-shaped topography. The stable meteorological conditions and the topography of the bell mouth in Xi’an make it difficult for the precursor pollutants to diffuse. These meteorological conditions and the concentration ratio of precursor pollutants in the areas surrounding Xi’an favor the occurrence of photochemical reactions. Therefore, the joint prevention and control of O₃ pollution in the Fenwei Plain must be strengthened in order to strictly control local pollution source emissions.

4. Conclusions

(1) From 2014 to 2020, the annual variation in the O₃_8h_max mass concentration in Xi’an shows a large interannual variability, rising from 65.58 μg/m³ in 2014 to 88.5 μg/m³ in 2020, with an increase of 34.94%. The monthly variation features an inverted “V” shape, rising from January, reaching the maximum value in June and July, and then showing a downward trend. The diurnal variation was significantly affected by the photochemical process of the near-surface atmosphere, presenting an obvious single-peak and single-valley distribution, and the O₃ concentrations were low in the morning and night. The variation trend of the mass concentration of NO₂, an important precursor pollutant for O₃ formation, was opposite to that of the O₃ in different seasons.

(2) The average mass concentration of O₃ was positively correlated with sunshine duration and surface temperature but had a complex correlation with the relative humidity and wind speed. With the increase in the sunshine duration and surface temperature, the mass concentration of O₃ increased correspondingly. However, with the increase in the relative humidity and wind speed, the average mass concentration of O₃ rose gently at first and then decreased rapidly. When the temperature was ≥30 °C, the relative humidity was 30~50%, and the wind speed was ≤4 m·s⁻¹, high concentrations of O₃ pollution easily occurred.

(3) The cluster analysis of the heavy O₃ pollution period in Xi’an from 2014 to 2020 shows that the air mass in Xi’an during the heavy O₃ pollution period was dominated by short-distance air flow, of which the short-range air flow in the northerly direction and the air flow in the southwest and northeast directions accounted for a large proportion.

(4) The analysis of the transport pathways and potential source during the heavy O₃ pollution period in Xi’an shows that the O₃ pollution in Xi’an was based on local sources and was superimposed on adjacent areas, such as the southwest of Northeast China, as well as the central part of Shaanxi Province, the southwest of Shanxi Province, and the northern part of Henan Province. A large number of O₃ precursors were jointly polluted, and the main pollution areas were concentrated in the Fenwei Plain. It is necessary to strengthen regional joint prevention and control to effectively curb the occurrence of O₃ pollution.