Next Article in Journal
Bioaerosol Seasonal Variation and Contribution to Airborne Particulate Matter in Huangshi City of Central China
Next Article in Special Issue
The Impact of Long-Range Transport of Biomass Burning Emissions in Southeast Asia on Southern China
Previous Article in Journal
Interference of Urban Morphological Parameters in the Spatiotemporal Distribution of PM10 and NO2, Taking Dalian as an Example
Previous Article in Special Issue
Evaluation of TROPOMI and OMI Tropospheric NO2 Products Using Measurements from MAX-DOAS and State-Controlled Stations in the Jiangsu Province of China
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Ozone Pollution in Chinese Cities: Spatiotemporal Variations and Their Relationships with Meteorological and Other Pollution Factors (2016–2020)

1
School of Computer and Information Engineering, Henan University, Kaifeng 475004, China
2
Henan Key Laboratory of Big Data Analysis and Processing, Kaifeng 475004, China
3
Henan Engineering Laboratory of Spatial Information Processing, Kaifeng 475004, China
*
Authors to whom correspondence should be addressed.
Atmosphere 2022, 13(6), 908; https://doi.org/10.3390/atmos13060908
Submission received: 30 April 2022 / Revised: 23 May 2022 / Accepted: 30 May 2022 / Published: 3 June 2022

Abstract

:
With the acceleration of urbanization, ozone (O3) pollution has become increasingly serious in many Chinese cities. This study analyzes the temporal and spatial characteristics of O3 based on monitoring and meteorological data for 366 cities and national weather stations throughout China from 2016 to 2020. Least squares linear regression and Spearman’s correlation coefficient were computed to investigate the relationships of O3 with various pollution factors and meteorological conditions. Global Moran’s I and the Getis–Ord index G i * were adopted to reveal the spatial agglomeration of O3 pollution in Chinese cities and characterize the temporal and spatial characteristics of hot and cold spots. The results show that the national proportion of cities with an annual concentration exceeding 160 μg·m−3 increased from 21.6% in 2016 to 50.9% in 2018 but dropped to 21.5% in 2020; these cities are concentrated mainly in Central China (CC) and East China (EC). Throughout most of China, the highest seasonal O3 concentrations occur in summer, while the highest values in South China (SC) and Southwest China (SWC) occur in autumn and spring, respectively. The highest monthly O3 concentration reached 200 μg·m−3 in North China (NC) in June, while the lowest value was 60 μg·m−3 in Northeast China (NEC) in December. O3 is positively correlated with the ground surface temperature (GST) and sunshine duration (SSD) and negatively correlated with pressure (PRS) and relative humidity (RHU). Wind speed (WIN) and precipitation (PRE) were positively correlated in all regions except SC. O3 concentrations are significantly differentiated in space: O3 pollution is high in CC and EC and relatively low in the western and northeastern regions. The concentration of O3 exhibits obvious agglomeration characteristics, with hot spots being concentrated mainly in NC, CC and EC.

1. Introduction

With the rapid development of industrialization and urbanization, surface ozone (O3) concentrations are rising in cities throughout China. In addition to the natural emissions of surface ozone, biological and anthropogenic volatile organic compounds (BVOCs and AVOCs, respectively) in the atmosphere have tremendous impacts on the formation of O3. BVOCs play a particularly important role globally, as their emission volume is estimated to be 10 times greater than that of AVOCs. Additionally, many BVOCs are characterized as highly reactive and can rapidly result in peroxy radicals, which are important for the formation of O3 and ultrafine particles [1,2,3,4,5,6]. As a consequence, after fine particulate matter (PM2.5), O3 has become one of the most harmful gases to urban air quality [7]. Specifically, surface O3 pollution can stimulate and damage human mucosal tissues (such as the respiratory system and eyes) and even damage the nervous and reproductive systems [8,9]. In addition, O3 increases agricultural losses and accelerates the formation of pollutants such as particulate matter, which in turn affects the frequency and intensity of high-pollution weather events [10,11]. Therefore, the rising O3 concentrations in many regions are drawing increasing concern.
Meteorological conditions further play a leading role in the diffusion, transformation and removal of atmospheric O3 [12,13,14,15]. Thus, if effective measures are to be implemented to control and reduce O3 concentrations, it is crucial to understand the temporal and spatial variations of O3 and its relationship with various pollution factors and meteorological conditions [16,17,18]. Accordingly, over the past ten years, many studies have investigated these issues.
Remote sensing data, model simulation data and ground station data are all commonly used to research air pollutants. For example, Zhang, JX. et al. [19] focused on the temporal and spatial dynamics of total O3 over China from 1981 to 2010 by using multisource remote sensing data and ground-level stations. Tan, K. C. et al. [20] used the Ozone Monitoring Instrument (OMI) on board the Aura satellite to obtain the total column ozone (TCO) and to examine the spatial and temporal distributions of atmospheric O3 over Peninsular Malaysia in 2015. Nevertheless, although remote sensing data have the advantages of a large detection range and few restrictions with regard to the ground conditions, the obtained data are the total O3 column concentrations, whereas the near-surface O3 concentration is often of interest. Hence, in addition to using remote sensing data, some studies have conducted modeling to simulate and study O3 concentrations. For example, Fu, Y. et al. [21] used the Model of Emissions of Gases and Aerosols from Nature (MEGAN) embedded within the global three-dimensional Goddard Earth Observing System Chemical Transport Model (GEOS-Chem) to simulate the concentrations of O3 in China between 2001 and 2006. Hu, J. et al. [22] conducted a year-long (2013) air quality simulation using the Weather Research and Forecasting (WRF) model and the Community Multiscale Air Quality (CMAQ) model to provide detailed spatiotemporal information on O3. However, although these model simulation data can solve the insufficient spatial coverage problem, the data generated by such modeling efforts are not convincing.
Surface O3 data were very sparse before 2013. Starting in 2013, the surface monitoring network greatly expanded, and the China Ministry of Ecology and Environment began to publicly provide detailed hourly data across all of China. These data comprise information necessary for assessing the temporal and spatial patterns of air quality in China and its driving factors. As a result, several studies have utilized these monitoring data to determine the spatiotemporal characteristics of O3 in China and its correlations with pollution and meteorological factors. For example, Zhang, Z. et al. [23] studied the temporal and spatial distributions of O3 and the influences of meteorological factors on O3 in the Yangtze River Delta region. Nevertheless, despite the long-term studies on O3 in the literature, most previous studies have been limited to urban agglomerations, whereas little work has been conducted on suburban or rural areas. Moreover, Mao, J. et al. [24] studied the relationships between O3 and meteorological factors in eastern China in 2017, but their analysis was limited to one year. In cases such as this, the temporal distribution of O3 cannot be fully explored.
Based on the above problems, this paper uses the hourly average concentrations of six pollutants (O3, nitrogen dioxide (NO2), carbon monoxide (CO), inhalable particulate matter (PM10), PM2.5 and sulfur dioxide (SO2)) in 366 cities in China and the daily averages of meteorological variables (including surface temperature, sunshine duration, pressure and relative humidity) from 839 national weather stations to study the spatiotemporal distribution of O3 during the period of the 13th Five-Year Plan (2016–2020). On this basis, the relationships between O3 and other air pollutants and meteorological factors are discussed, as are the effects of other air pollutants and meteorological factors on O3, and the spatial aggregation characteristics of O3 are analyzed.

2. Materials and Methods

2.1. Study Area

Considering geography, climate, economy and other factors, this article divides China into seven regions (Figure 1) according to the Chinese geographic regional classification scheme: Central China (CC), East China (EC), North China (NC), South China (SC), Northwest China (NWC), Southwest China (SWC), and Northeast China (NEC) [25].

2.2. Dataset

2.2.1. Air Quality Data

The data analyzed in this study include hourly O3 concentrations recorded at China Environmental Monitoring Stations in 366 cities across China. The start and end times of the data are 1 January 2016, and 31 December 2020, respectively. The study area encompasses all of China, including all prefecture-level cities (except Hong Kong, Macao, and Taiwan).
According to the O3 concentration specifications and standards of the Ambient Air Quality Standard (GB3095-2012), the original data were subjected to quality control according to the data validity requirements, and negative hourly O3 concentrations and missing values in the original data were removed. The seasonal average refers to the arithmetic average of the maximum 8 h average concentrations on each day during a given season, while the annual average refers to the arithmetic average of the maximum 8 h average concentrations on each day in that calendar year.

2.2.2. Meteorological Data

The acquired meteorological data include the China Surface Climate Data Daily Value Dataset (V3.0) and Ground Cumulative Annual Value Dataset published by the China Meteorological Science Data Sharing Network (http://data.cma.cn, accessed on 30 April 2022). The meteorological elements considered include the ground surface temperature (GST), pressure (PRS), relative humidity (RHU), wind speed (WIN), sunshine duration (SSD) and precipitation (PRE), the data of which originated from 699 stations across China. The original data files were strictly controlled and inspected to exclude outliers and null values from the dataset.

2.3. Methodology

2.3.1. Evaluation Standard

The newly implemented Ambient Air Quality Evaluation Technical Specification (Trial) (HJ663-2013) in China requires that the daily evaluation should consider the maximum 8 h average O3 concentration in each city to determine its daily compliance status, whereas the monthly evaluation should consider the maximum daily O3 concentrations in a given month in each city (here, the 90th percentile of the 8 h moving average is used to determine monthly compliance), and the annual evaluation depends on the 90th percentile of the highest 8 h moving average value of the daily O3 concentrations throughout the year to determine the annual compliance status.
O3 concentration limits are associated with different standards according to different functional areas. The first-level excess concentration threshold is 100 µg·m−3, and the second-level excess concentration threshold is 160 µg·m−3. This article uniformly applies the second-class environmental function zone standard to determine whether the daily, monthly and annual average values in each city satisfy the national standards.

2.3.2. Multiple Linear Regression

In regression analysis, if there are two or more independent variables, the process is called multiple regression. In the real world, the target variable is often associated with multiple variables, so the target variable is jointly estimated by the optimal combination of multiple independent variables. Multiple linear regression is an extension of univariate linear regression. Its basic formula is as follows:
y = a x 1 + b x 2 + + n x m + ε
This paper uses multiple linear regression to test the impacts of other air pollutants on O3. A larger coefficient corresponding to a certain pollutant signifies that that pollutant can contribute more to the generation of O3.

2.3.3. Global Moran’s I Index

Since near-ground O3 is continuously distributed in space and because atmospheric photochemical reaction precursors in cities across the country interact with each other, this article applies global Moran’s I index [26] to test the spatial autocorrelation of O3 on a national scale [27,28]. The index is calculated as follows:
I   = n S 0 i   = 1 n j   = 1 n w ij Z i Z j i   = 1 n Z i 2
where Z i is the deviation between the attribute value for element i and the average value (xi X ¯ ); w i , j denotes the spatial weights between elements i and j; n is the total number of elements; and
S 0 = i   = 1 n j   = 1 n w i , j  
Global Moran’s I was tested for significance using the Z test, and the calculation formula is
Z [ I ] = I E [ I ] V [ I ]
where Z [ I ] is the Z test value of global Moran’s I; E [ I ] is the mathematical expectation; and V [ I ] is the variance
Generally, I is between −1.0 and 1.0. When I ∈ [−1, 0), there is a negative spatial correlation. The closer the value of I is to −1, the greater the difference in attribute values between spatial units. When I = 0, the spatial distribution is random. As I approaches 0, the attribute values between spatial units become increasingly irrelevant. When I ∈ (0, 1], the correlation is spatially positive, and values closer to 1 indicate a stronger correlation of attribute values between spatial units.

2.3.4. Getis–Ord G i * Index

Although global Moran’s I index can reflect the spatial correlation of O3 concentrations as a whole, it cannot reflect the sign and degree of correlation between any two cities. The Getis–Ord index G i * [29,30] is selected for a local area to judge the spatial heterogeneity in the area by analyzing local information, revealing cold spots and hot spots of O3 concentrations that reflect the degree of correlation between neighboring areas. This index is calculated as follows:
G i * = j   = 1 n w i , j x j   X ¯ j   = 1 n w i , j S [ n j   = 1 n w i , j 2     ( j   = 1 n w i , j ) 2 ] n 1
where xj is the attribute value for element j; w i , j is the spatial weight matrix between elements i and j; n is the total number of elements; and
X ¯   = j   = 1 n x j n
S   = j   = 1 n x j 2 n ( X ¯ ) 2
The G i * statistic is the Z score. If Z( G i * ) is significant and positive, it indicates that the value near position i is higher than the average, and the higher the Z( G i * ) score is, the greater the number of high-value clusters (hot spots). In contrast, if Z( G i * ) is significant and negative, it indicates that the value near position i is lower than the average, and the lower the Z( G i * ) score is, the greater the number of low-value clusters (cold spots). This article uses the Hot Spot Analysis (Getis–Ord G i * ) tool for ArcGIS 10.7 software (Esri, California, United States) to analyze the annual, quarterly and monthly O3 concentrations in 366 cities in China and to conduct an analysis of hot and cold spots.

3. Analysis of Spatiotemporal Characteristics and Influencing Factors

3.1. Analysis of Temporal and Spatial Changes

3.1.1. Temporal and Spatial Changes in the Annual Average Ozone Concentration

From 2016 to 2020, the average 90th percentile values of the 8 h average O3 concentrations in Chinese cities were (successively) 138.7, 149.3, 151.4, 147.1 and 126.8 µg·m−3, and the O3 concentrations exhibited similar variations in each year. Among the regions, due to the large number of cities in CC, EC and SC, human factors appear to have certain impacts on the O3 concentration and caused the O3 concentration to rise until 2019 [31]. In contrast, the number of cities in NWC is relatively small, so the O3 concentration began to gradually decline in 2017. NC, SWC and NEC have moderate numbers of cities, so these regions show the same trend as the whole country; that is, the O3 concentrations rose up until 2019, whereupon they started to decline.
The annual average O3 concentrations have strong spatial characteristics (Figure 2). Over the past five years, most cities with O3 concentrations exceeding 160 μg·m−3 were located in the southern part of NC and the northern parts of CC and EC, and the proportion increased from 20% in 2016 to 50% in 2018. This percentage finally declined to 20% in 2020. The cities with O3 concentrations ranging from 100 to 160 μg·m−3 were evenly distributed throughout China, and the proportion remained 50% or higher. In addition, most cities with O3 concentrations below 100 μg m−3 were located in NEC and SWC, and the proportion remained 5% or lower.

3.1.2. Temporal and Spatial Changes in the Seasonal Average Ozone Concentration

The spatial patterns of O3 concentrations exhibited substantial seasonal variations (Figure 3), according to which China could be divided into three regions. The first region includes NC, NWC and NEC. The O3 concentrations in this region were the highest in summer and the lowest in winter. The concentrations in spring and autumn were moderate, and the O3 concentrations in spring were higher than those in autumn, especially in NEC. The second region includes CC and EC. The summer and winter O3 concentrations in this region were comparable to those in the first region and were similar to the spring and autumn O3 concentrations. The third area includes SC and SWC. The seasonal O3 concentrations in this area were relatively close, but the O3 concentrations remained the lowest in winter. However, in SC, the O3 concentrations in autumn were the highest, while in SWC, the O3 concentrations in spring were the highest. From north to south, the absolute difference in the O3 concentration among the four seasons gradually decreased. Due to climatic reasons, the absolute temperature difference among the four seasons is large in the north but small in the south, which is warm year-round [32]. Hence, temperature affects the rate at which precursors are converted into O3.

3.1.3. Temporal and Spatial Changes in the Monthly Average Ozone Concentration

The monthly O3 concentration was spatially correlated, and the number of cities exceeding the standard in each region presented different variation characteristics (Figure 4). Due to the large number of urban agglomerations in the eastern regions, the number of cities in EC exceeding the standard in the whole year was higher than that in the western regions. The higher latitudes of the northern Chinese regions lead to an earlier onset of warming each year, so the northern regions reached the highest number of cities exceeding the standard earlier than the regions to their south. In addition, due to the large temperature difference across the northern regions throughout the year, the number of cities exceeding the standard varied greatly from month to month. For example, the number of cities in EC exceeding the standard in May was 343 times that in January. In contrast, because the latitudes and longitudes of the southern Chinese regions are relatively low, the temperature rises slowly, and the temperature difference is small throughout the year; consequently, only a small number of cities exceeded the standard among the southern regions throughout the year, and the temperature distribution was relatively uniform each month.
During January–March, the number of cities exceeding the standard was relatively small; examples of these cities include Suzhou in NC in 2016, Beihai in SC in 2017 and Guoluo in SWC in 2019. With a gradual increase in temperature, the speed with which O3 is generated increases, and the number of cities exceeding the standard begins to increase [32]. Among the different regions, EC had a particularly high number of cities exceeding the standard. From April to September, the number of such cities accounted for approximately 50% of the national total and peaked at 343 in May, 15.5 times the number in SC. Meanwhile, the numbers of cities exceeding the standard in CC and NC reached approximately 20% of the national total and peaked at 135 (September) and 150 (June), respectively. In NEC, NWC and SWC, on the one hand, there are relatively few urban agglomerations, and on the other hand, the temperature is relatively low; as a result, few cities exceeded the standard (only 23, 9, and 3 cities, respectively, exceeded the standard in September). In addition, the changes in the number of cities exceeding the standard in SC were relatively unique; due to the low latitude, the temperature rises slowly, and the change range is small, causing the number of cities exceeding the standard to peak at 61 in September. Over the next 10–12 months, only the three southernmost regions (CC, EC and SC) contained cities with excess O3, but their numbers were slowly decreasing. Among them, SC (the southernmost tip of China) has always maintained a high level; nevertheless, even in December, only two cities in SC, Zhuhai and Yangjiang, exceeded the standard, with concentrations of only 161 μg·m−3 and 163 μg·m−3, respectively.

3.2. Analysis of the Relationships between O3 and Other Pollution Factors and Changes in Hourly Concentrations

The hourly concentration trends of O3 and five other gases were compared (Figure 5). Because the change trend of O3 is clearly related to those of these other gases, their correlations were tested by using Spearman’s correlation coefficient and the T test. Since the significance and correlation coefficients of the linear regression equations in each region are relatively similar, here, only the results for all of China are discussed.
The calculated correlation coefficient between NO2 and O3 is −0.753, the T value is −108.97, and the Pt value is 0, indicating that these two gases attain a significant negative correlation. Atmospheric O3 at ground level is formed in the presence of ultraviolet (UV) light (λ < 424 nm) through the direct photolysis of NO2 and VOCs ( N O 2 + V O C s   U V _ _     O 3 ) [3,33], and sunlight accelerates the photochemical conversion of NO2 into O3, causing O3 concentrations to increase and NO2 concentrations to decrease during the day; in contrast, at night, the rate of this reaction is slow, resulting in relatively low O3 concentrations.
CO attains a correlation coefficient of −0.791 with O3, and the T value is −85.33, while the Pt value is 0. This indicates that these two gases also achieve a significant negative correlation. Under UV irradiation, CO reacts with oxygen to form O3 and carbon dioxide ( C O + 2 O 2   U V _ _     C O 2 + O 3 ) [34], and sunlight further increases the rate of this photochemical reaction, increasing O3 concentrations and decreasing CO concentrations during the day; conversely, the slow photochemical reaction rate at night results in relatively low O3 concentrations.
PM2.5 and O3 attain a correlation coefficient of −0.794 with a T value of −66.31 and a Pt value of 0, similarly indicating a significant negative correlation between these two pollutants. The reason for the low correlation coefficient is that PM2.5 and O3 exhibit a coupling mechanism. When the value of the kinetic parameter, i.e., the hydrogen superoxide (HO2) uptake coefficient, is 0.2, the change in the heterogeneous HO2 uptake rate exerts an important impact on O3 [35], while a reduction in PM2.5 leads to an increase in HO2, which increases the number of OH radicals. These OH radicals then generate O3 under photochemical reactions, causing the opposing trends of PM2.5 and O3.
In addition to utilizing the Spearman correlation coefficient to verify the effects of other air pollutants on O3, a multiple regression analysis was also performed. A larger parameter for a certain pollutant suggests that it has a greater impact on the O3 concentration. The expression obtained after multiple iterations of training is as follows (8):
  O 3     ( 0.744 ) C O + ( 0.38 ) N O 2 + ( 0.442 ) P M 2.5 + ( 0.017 ) P M 10 + ( 0.08 ) S O 2
The approximately equal sign is used in this formula because the overall error is not considered here. Nevertheless, the parameters of NO2, CO and PM2.5 are clearly larger than those of PM10 and SO2, which means that the former three air pollutants have greater impacts on O3.

3.3. Analysis of the Relationships between O3 and Meteorological Factors

Upon testing the significance of the correlation of certain meteorological factors with O3, some factors attained a p value greater than 0.01, indicating that the correlation is not significant [36]. These meteorological factors are not examined here. It follows, then, that the correlations discussed below are all significant; i.e., their P values are less than 0.01. In addition, we stipulate that the correlation can be explained only when the correlation coefficient |r| > 0.3 [37].
The annual average O3 concentrations in China exhibited significant spatial correlations with GST, PRS, RHU, WIN, SSD and PRE during 2016–2020 (Figure 6). The colored bars extending above and below the invisible axis in each panel of the figure signify positive and negative correlation coefficients, respectively. GST and SSD attain a significant positive correlation in all regions. This is because high temperatures and UV radiation speed up the rate at which reactions with molecules such as NO2 produce O3 [38,39]. PRS exhibits a significant negative correlation in most regions because O3 is easily soluble in water, and an increase in PRS accelerates the rate at which O3 dissolves in water [40,41]. RHU achieves a significant negative correlation in all regions (except NC) because RHU affects UV exposure and reduces the rate at which other air pollutants produce O3 [42,43]. WIN attains a significant negative correlation in CC and SC but a significant positive correlation in the remaining regions. The reason for this phenomenon is that high wind speeds lift the height of the atmospheric boundary layer, causing the O3 in the upper layer to be mixed down to the surface, resulting in an increase in the concentration of near-surface O3 [44,45]. Finally, except for the significant negative correlation attained in SC, PRE exhibits a significant positive correlation in the other regions. This means that large amounts of PRE reduce O3 concentrations, while small amounts of PRE increase O3 concentrations. When there is a large amount of PRE, the cloud cover is generally greater, and the cloud layer—which absorbs UV rays—reduces the rate of O3 generation [46,47]. In addition, O3 dissolves into the water film in large quantities, resulting in a decrease in the O3 concentration [48].
The seasonal average O3 concentrations in China and the seasonal averages of GST, PRS, RHU, WIN, SSD and PRE exhibited notable spatial correlations during 2016–2020 (Figure 6). According to the data, GST is basically positively correlated with O3 in all regions. GST exerts a greater impact during spring and autumn than during summer and winter and in the eastern and northern regions than in the western and southern regions. On the one hand, this variability is due to the large temperature changes that occur in spring and autumn, which shows a clear correlation. On the other hand, compared with the south, the north experiences greater temperature changes in the four seasons, so the correlation in the north is higher than that in the south [38,39].
In terms of PRS, because the changing of seasons has less effect on PRS, PRS is negatively correlated to a similar degree with O3 in all seasons. PRS further increases the water solubility of O3, causing O3 concentrations to decrease with increasing PRS [40,49]. Likewise, RHU is negatively correlated with O3 in all seasons. Since O3 is easily soluble in water, O3 concentrations gradually decrease with increasing RHU [41].
In contrast, WIN displays a positive correlation with O3 in all regions during all four seasons of the year. Among the various regions, the influence of the wind speed in the north is obviously greater than that in the south because there are many plains in the north, and the wind speed is often higher there than in the south [42,43]. Similarly, SSD is positively correlated with O3 throughout the year, and the correlation coefficients in the south are higher than those in the north. Because SSD directly reflects the radiation intensity, a high level of irradiation accelerates the conversion of O3 precursors into O3, thereby increasing the O3 concentrations [44,45].
Finally, PRE is negatively correlated with the seasonal average O3 concentrations in China in summer, while in the other seasons, PRE generally attains a negative correlation in the south while typically reaching a positive correlation in the north. This is because when PRE is weak, O3 is brought from the stratosphere close to the surface, resulting in higher near-surface O3 levels. In contrast, when PRE is heavy, the humidity further increases, causing more O3 to dissolve in the water and thus reducing the O3 concentrations [46,47].

4. Spatial Agglomeration Characteristics of O3 Concentrations in Chinese Cities

4.1. Spatial Autocorrelation Test of O3 Concentrations

The spatial autocorrelation (Moran’s I) index of the annual mean, quarterly mean and monthly mean O3 concentrations were calculated for all 366 cities by the ArcGIS Spatial Analyst tool. Generally, if Z[I] > 1.65 or Z[I] < −1.65, the confidence is 90%; if Z[I] > 1.96 or Z[I] < −1.96, the confidence is 95%; and if Z[I] > 2.58 or Z[I] < −2.58, the confidence is 99%. When Z[I] < −2.58, the spatial distribution of the O3 concentration is negatively correlated and scattered; conversely, when Z[I] > 2.58, the O3 concentration is positively correlated in space (i.e., high values agglomerate in separate clusters, as do low values, forming distinct hot and cold spots).
As shown in Table 1 and Table 2, Moran’s I indexes are all positive, and all Z[I] values are greater than 2.58, passing the 1% test of significance and indicating that China’s urban O3 concentrations during 2016–2020 were spatially autocorrelated. Each region is assigned a single city as the basic unit. In addition to the different characteristics between different regions, there are also differences in the natural environmental conditions of different cities within the same region, which together lead to changes in the air quality in each region. Therefore, cluster analysis was performed on each city to find the hot and cold spots (i.e., where O3 is concentrated) to better understand the mechanisms of O3 pollution in different regions and to better formulate targeted prevention and control policies.

4.2. Spatial Agglomeration Characteristics of O3 Concentrations

The average Z[I] index for the past five years is 41.15, indicating that the O3 concentrations are spatially concentrated. Figure 7 shows that hot spots are concentrated mainly in NC, southern NEC, eastern NWC, northern CC, and central and northern EC; in contrast, cold spots are concentrated mainly in northern Xinjiang, northern NEC, central and southern SWC, and SC. However, the spatial autocorrelation of O3 is not obvious in various scattered cities throughout central NC, central and southern NWC, CC, southern EC and northern SWC.
In recent years, with the advancement of market-oriented reforms, the urbanization and industrialization levels of some cities in EC and NC have continuously risen. As some of these cities play important roles in driving economic development, in the development of these economies, automobile exhaust and industrial waste gas are continuously and inevitably emitted. Hence, the emission of man-made precursors raises the O3 concentrations in these urban agglomerations relative to those in surrounding cities.
Meteorological conditions, especially temperature, have an important influence on the O3 concentrations near the ground. In different seasons, changes in the solar radiation intensity lead to large seasonal differences in the concentrations of O3, and periodic changes are evident (Figure 7). In spring, hotspot cities are distributed mainly in southern NEC, NC, northern CC and EC. The distribution of hotspot cities in summer is further expanded and ranges from the east coast to inland areas, indicating that the photochemical reactions of O3 precursors caused by solar radiation are enhanced in summer. The extent of hotspot cities further increases in autumn and spreads to the southeast coast, with these cities concentrated mainly in southern NC, EC, CC and SC. Then, in winter, the breadth of hotspot cities shrinks sharply but continues to expand south, with these cities appearing mainly in southern SWC, southern EC and SC. The cold spot cities in spring and summer are concentrated mainly in NWC, SWC and SC; those in autumn are concentrated in NWC, SWC and NEC; and those in winter are concentrated predominantly in NEC, NC, northern CC and EC. Except for some cities in the northern parts of NWC and NEC, no area in the country features stable and continuously low O3 concentrations.

5. Conclusions

This paper analyzes the temporal and spatial changes in urban O3 pollutants across China from 2016 to 2020, as well as the influencing factors and spatial agglomeration characteristics of O3.
(1)
The O3 concentrations exhibited the same change trend year over year. However, the O3 concentrations in major urban agglomerations continued to rise in 2019. Seasonally, O3 has strong spatial characteristics and can be latitudinally divided into three regions from south to north. On a monthly scale, many cities exhibited O3 concentrations exceeding the national standard (160 μg·m−3) from April to September, with the peak appearing in June.
(2)
Cities with heavy O3 pollution were concentrated in the northern part of CC, southern NC and EC. Over time, the number of cities with O3 concentrations above 160 μg·m−3 has gradually increased, with the highest value being discovered in 2018, accounting for 50% of the total number of cities. In terms of monthly O3 concentrations, the number of cities with O3 concentrations above 160 μg·m−3 was relatively high from April to September, and most of these cities were concentrated in CC, EC and NC.
(3)
O3 displays significant negative correlations with NO2, PM2.5 and PM10, and its correlation coefficient with NO2 is the highest (with an r value of −0.399). The correlations between O3 and various meteorological factors vary both seasonally and spatially. GST and SSD show a significant positive correlation with O3 in all regions. PRS and RHU show significant negative correlations in most regions, while WIN and PRE show significant positive correlations in most regions. The effects of GST and PRS are greater in spring and autumn than in summer and winter. PRE is negatively correlated with the O3 concentration in summer and negatively correlated in the other seasons. For RHU, WIN and SSD, their correlations with O3 do not significantly differ from season to season.
(4)
The concentrations of O3 in Chinese cities have significant spatial agglomeration characteristics. Seasonal differences in the O3 concentration are large, and periodic changes are obvious. The hotspot cities in spring are distributed mainly in the southern part of NEC, NC, EC and northern CC. The distribution of summer hotspot cities further expands from the east coast to the inland areas. In autumn, the scope of hotspot cities expands further and spreads to the southeast coast, mainly in the southern part of NC, EC, CC and SC. The scope of winter hotspot cities shrinks sharply but continues to expand south, with cities appearing mainly in SWC, SC, EC and SC.
Due to the complex formation mechanism of O3 pollution, many factors, such as the concentrations of precursors, emission intensity, and meteorological conditions, can affect O3. Accordingly, the interactions between the O3 concentration and its various influencing factors need to be further studied. Only by fully revealing the causes and sources of O3 in different regions and quantifying the leading factors influencing the formation of O3 can we effectively improve the prevention and control of O3 pollution. Therefore, the state’s prevention and control of O3 can be carried out from two aspects. On the one hand, it is necessary to treat O3 precursors (NOx and VOCs) at the same time to reduce the generation of O3. On the other hand, controlling meteorological changes (for example, introducing artificial rainfall in summer when the temperature is higher) can reduce the concentrations of O3.

Author Contributions

Conceptualization, Q.G.; methodology, K.C.; formal analysis, X.Z.; data curation, K.C.; writing—original draft preparation, X.Z.; writing—review and editing, Q.G. and K.C.; visualization, X.Z.; funding acquisition, Y.L. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant Nos. 42071409 and 62176087) and Key Research and Promotion Projects of Henan Province (Grant No. 212102210079).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

Air quality data and meteorological data are available from http://data.cma.cn, accessed on 29 August 2021.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Najafpoor, A.; Jonidi, A.A.; Dousti, S. Trend Analysis of Air Quality Index Criteria Pollutants (Co, NO2, SO2, PM10 and O3) Concentration Changes in Tehran Metropolis and Its Relationship with Meteorological Data, 2001–2009. J. Health Field 2017, 3, 17–26. [Google Scholar]
  2. Nwaogazie, I.L.; Wilson, A.H.; Henshaw, T. Modeling the Effect of Atmospheric Stability, Nitrogen Oxide and Carbon Monoxide on the Formation of Ozone: A Case of Ogba-Egbema-Ndoni Local Government Area in Nigeria. Int. J. Civ. Eng. 2016, 7, 111–121. [Google Scholar]
  3. Qu, Z.; Henze, D.K.; Cooper, O.R.; Neu, J.L. Impacts of Global Nox Inversions on NO2 and Ozone Simulations. Atmos. Chem. Phys. 2020, 20, 13109–13130. [Google Scholar] [CrossRef]
  4. Stohl, A.; Williams, E.; Wotawa, G.; Kromp-Kolb, H. A European Inventory of Soil Nitric Oxide Emissions and the Effect of These Emissions on the Photochemical Formation of Ozone. Atmos. Environ. 1996, 30, 3741–3755. [Google Scholar] [CrossRef]
  5. Wang, P.; Chen, Y.; Hu, J.; Zhang, H.; Ying, Q. Attribution of Tropospheric Ozone to No_X and Voc Emissions: Considering Ozone Formation in the Transition Regime. Environ. Sci. Technol. 2019, 53, 1404–1412. [Google Scholar] [CrossRef] [PubMed]
  6. Kaser, L.; Peron, A.; Graus, M.; Striednig, M.; Wohlfahrt, G.; Juráň, S.; Karl, T. Interannual Variability of Terpenoid Emissions in an Alpine City. Atmos. Chem. Phys. 2022, 22, 5603–5618. [Google Scholar] [CrossRef]
  7. Ordonez, C.; Garrido-Perez, J.M.; Garcia-Herrera, R. Early Spring near-Surface Ozone in Europe During the COVID-19 Shutdown: Meteorological Effects Outweigh Emission Changes. Sci. Total Environ. 2020, 747, 141322. [Google Scholar] [CrossRef]
  8. Alexis, N.E.; Lay, J.C.; Hazucha, M.; Harris, B.; Hernandez, M.L.; Bromberg, P.A.; Kehrl, H.; Diaz-Sanchez, D.; Kim, C.; Devlin, R.B.; et al. Low-Level Ozone Exposure Induces Airways Inflammation and Modifies Cell Surface Phenotypes in Healthy Humans. Inhal. Toxicol. 2010, 22, 593–600. [Google Scholar] [CrossRef] [Green Version]
  9. Lefohn, A.S.; Malley, C.S.; Smith, L.; Wells, B.; Gerosa, G. Tropospheric Ozone Assessment Report: Global Ozone Metrics for Climate Change, Human Health, and Crop/Ecosystem Research. Elem. Sci. Anthr. 2018, 6, 28. [Google Scholar] [CrossRef] [Green Version]
  10. Faridi, S.; Shamsipour, M.; Krzyzanowski, M.; Kunzli, N.; Amini, H.; Azimi, F.; Malkawi, M.; Momeniha, F.; Gholampour, A.; Hassanvand, M.S.; et al. Long-Term Trends and Health Impact of Pm2.5 and O-3 in Tehran, Iran, 2006–2015. Environ. Int. 2018, 114, 37–49. [Google Scholar] [CrossRef]
  11. Castell, J.F. Impacts De Lozone Sur Lagriculture Et Les Forts Et Estimation Des Cots Conomiques Ozone Impacts on Agriculture and Forests and Economic Losses Assessment. Pollut. Atmosphérique 2019, 229–230, 142–152. [Google Scholar]
  12. Flaum, J.B.; Rao, S.T.; Zurbenko, I.G. Moderating the Influence of Meteorological Conditions on Ambient Ozone Concentrations. J. Air Waste Manag. Assoc. 1996, 46, 35–46. [Google Scholar] [CrossRef] [PubMed]
  13. Liu, P.; Song, H.; Wang, T.; Wang, F.; Zhao, H. Effects of Meteorological Conditions and Anthropogenic Precursors on Ground-Level Ozone Concentrations in Chinese Cities. Environ. Pollut. 2020, 262, 114366. [Google Scholar] [CrossRef] [PubMed]
  14. Tong, N.Y.O.; Leung, D.Y.C.; Liu, C.H. A Review on Ozone Evolution and Its Relationship with Boundary Layer Characteristics in Urban Environments. Water Air Soil Pollut. 2011, 214, 13–36. [Google Scholar] [CrossRef]
  15. Wu, J.; Li, C.; Ma, Z.Q.; Sun, Z.B.; Dong, F. Influence of Meteorological Conditions on Ozone Pollution at Shangdianzi Station Based on Weather Classification. Huan Jing Ke Xue = Huanjing Kexue 2020, 41, 4864–4873. [Google Scholar] [PubMed]
  16. Fan, L.P.; Fu, S.; Wang, X.; Fu, Q.Y.; Jia, H.H.; Xu, H.; Qin, G.M.; Hu, X.; Cheng, J.P. Spatiotemporal Variations of Ambient Air Pollutants and Meteorological Influences over Typical Urban Agglomerations in China during the COVID-19 Lockdown. J. Environ. Sci. 2021, 106, 26–38. [Google Scholar] [CrossRef] [PubMed]
  17. Li, L.; Wang, K.; Li, S.; Feng, X.; Zhang, L. Lst-Net: Learning a Convolutional Neural Network with a Learnable Sparse Transform. In European Conference on Computer Vision; Springer: Cham, Switzerland, 2020; pp. 562–579. [Google Scholar]
  18. Wang, Z.B.; Li, J.X.; Liang, L.W. Spatio-Temporal Evolution of Ozone Pollution and Its Influencing Factors in the Beijing-Tianjin-Hebei Urban Agglomeration. Environ. Pollut. 2020, 256, 113419. [Google Scholar] [CrossRef]
  19. Zhang, J.X.; Zhang, Y.; Zhu, S.Y. Variation of Total Ozone over China for 30 Years Analyzed by Multi-Source Satellite Remote Sensing Data. Geo. Spat. Inf. Sci. 2014, 16, 971–978. [Google Scholar]
  20. Tan, K.C.; Lim, H.S.; Jafri, M.Z. Satellite Remote Sensing of Total Column Ozone over Peninsular Malaysia. Presented at the 6th IEEE International Conference on Control System, Computing and Engineering, Penang, Malaysia, 25–27 September 2016; pp. 374–379. [Google Scholar]
  21. Fu, Y.; Liao, H. Simulation of the Interannual Variations of Biogenic Emissions of Volatile Organic Compounds in China: Impacts on Tropospheric Ozone and Secondary Organic Aerosol. Atmos. Environ. 2012, 59, 170–185. [Google Scholar] [CrossRef]
  22. Hu, J.; Chen, J.; Ying, Q.; Zhang, H. One-Year Simulation of Ozone and Particulate Matter in Chinausing Wrf/Cmaq Modeling System. Atmos. Chem. Phys. 2016, 16, 10333–10350. [Google Scholar] [CrossRef] [Green Version]
  23. Zhang, Z.; Yao, M.; Wu, W.; Zhao, X.; Zhang, J. Spatiotemporal Assessment of Health Burden and Economic Losses Attributable to Short-Term Exposure to Ozone During 2015–2018 in China. J. Atmos. Sci. 2020, 21, 1–18. [Google Scholar]
  24. Mao, J.; Wang, L.; Lu, C.; Liu, J.; Wang, Y. Meteorological Mechanism for a Large-Scale Persistent Severe Ozone Pollution Event over Eastern China in 2017. J. Environ. Sci. 2020, 92, 187–199. [Google Scholar] [CrossRef] [PubMed]
  25. Zhuang, Y.; Li, R.Y.; Yang, H.; Chen, D.L.; Chen, Z.Y.; Gao, B.B.; He, B. Understanding Temporal and Spatial Distribution of Crop Residue Burning in China from 2003 to 2017 Using Modis Data. Remote Sens. 2018, 10, 390. [Google Scholar] [CrossRef] [Green Version]
  26. Geary, R.C. The Contiguity Ratio and Statistical Mapping. Inc. Stat. 1954, 5, 115–146. [Google Scholar] [CrossRef]
  27. Goodchild, M.; Haining, R.; Wise, S. Integrating Gis and Spatial Data Analysis: Problems and Possibilities. Int. J. Geogr. Inf. Syst. 1992, 6, 407–423. [Google Scholar] [CrossRef]
  28. Moran, P.A.P. The Interpretation of Statistical Maps. J. R. Stat. Soc. Ser. A Stat. Soc. 1948, 10, 243–251. [Google Scholar] [CrossRef]
  29. Ord, J.K. Local Spatial Autocorrelation Statistics: Distributional Issues and an Application. Geogr. Anal. 1995, 27, 286–306. [Google Scholar] [CrossRef]
  30. Getis, A. Spatial Interaction and Spatial Autocorrelation: A Cross-Product Approach. Environ. Plan. A 2010, 23, 1269–1277. [Google Scholar] [CrossRef]
  31. Gong, K.; Li, L.; Li, J.; Qin, M.; Wang, X.; Ying, Q.; Liao, H.; Guo, S.; Hu, M.; Zhang, Y.; et al. Quantifying the Impacts of Inter-City Transport on Air Quality in the Yangtze River Delta Urban Agglomeration, China: Implications for Regional Cooperative Controls of Pm2.5 and O3. Sci. Total Environ. 2021, 779, 146619. [Google Scholar] [CrossRef]
  32. Varotsos, K.V.; Giannakopoulos, C.; Tombrou, M. Ozone-Temperature Relationship During the 2003 and 2014 Heatwaves in Europe. Reg. Environ. Chang. 2019, 19, 1653–1665. [Google Scholar] [CrossRef]
  33. Tang, G.; Li, X.; Wang, Y.; Xin, J.; Ren, X. Surface Ozone Trend Details and Interpretations in Beijing, 2001–2006. Atmos. Chem. Phys. 2009, 9, 8813–8823. [Google Scholar] [CrossRef] [Green Version]
  34. Parrish, D.D.; Trainer, M.; Holloway, J.S.; Yee, J.E.; Warshawsky, M.S.; Fehsenfeld, F.C.; Forbes, G.L.; Moody, J.L. Relationships between Ozone and Carbon Monoxide at Surface Sites in the North Atlantic Region. J. Geophys. Res. 1998, 103, 13357–13376. [Google Scholar] [CrossRef]
  35. Li, Z.; Yu, S.C.; Wang, L.Q.; Mehmood, K.; Liu, W.P. Suppression of Convective Precipitation by Elevated Man-Made Aerosols Is Responsible for Large-Scale Droughts in North China. Proc. Natl. Acad. Sci. USA 2018, 115, 8327–8328. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  36. Kraemer, H.C.; Bloch, D.A. A Goodness-of-Fit Approach to Inference Procedures for the Kappa Statistic: Confidence Interval Construction, Significance-Testing and Sample Size Estimation. Stat. Med. 2010, 11, 1511–1519. [Google Scholar] [CrossRef] [PubMed]
  37. Lin, I.K. A Concordance Correlation-Coefficient to Evaluate Reproducibility. Biometrics 1989, 45, 255–268. [Google Scholar] [CrossRef] [PubMed]
  38. Zhou, X.L.; Liang, J.Q. Pollution Characteristics of Ozone in Urban Area of Foshan and Correlation between Ozone and Meteorological Factors. Environ. Monit. Assess. 2016, 8, 39–44. [Google Scholar]
  39. Sun, W.; Hess, P.; Liu, C. The Impact of Meteorological Persistence on the Distribution and Extremes of Ozone. Geophys. Res. Lett. 2017, 44, 1545–1553. [Google Scholar] [CrossRef]
  40. Zhan, C.; Xie, M.; Liu, J.; Wang, T.; Li, M. Surface Ozone in the Yangtze River Delta, China: A Synthesis of Basic Features, Meteorological Driving Factors and Health Impacts. J. Geophys. Res. Atmos. 2021, 126, e2020JD033600. [Google Scholar] [CrossRef]
  41. Chen, Z.; Zhuang, Y.; Xie, X.; Chen, D.; Cheng, N.; Yang, L.; Li, R. Understanding Long-Term Variations of Meteorological Influences on Ground Ozone Concentrations in Beijing During 2006–2016. Environ. Pollut. 2019, 245, 29–37. [Google Scholar] [CrossRef]
  42. Kumari, S.; Jayaraman, G.; Ghosh, C. Analysis of Long-Term Ozone Trend over Delhi and Its Meteorological Adjustment. Int. J. Environ. Sci. Technol. 2013, 10, 1325–1336. [Google Scholar] [CrossRef] [Green Version]
  43. Huang, L.S.; Smith, R.L. Meteorologically-Dependent Trends in Urban Ozone. Environmetrics 2015, 10, 103–118. [Google Scholar] [CrossRef]
  44. Chen, Z.; Li, R.; Chen, D.; Zhuang, Y.; Li, M. Understanding the Causal Influence of Major Meteorological Factors on Ground Ozone Concentrations across China. J. Clean. Prod. 2019, 242, 118498. [Google Scholar] [CrossRef]
  45. Bei, N.; Lei, W.; Zavala, M.; Molina, L.T. Ozone Predictabilities Due to Meteorological Uncertainties in the Mexico City Basin Using Ensemble Forecasts. Atmos. Chem. Phys. 2010, 10, 6295–6309. [Google Scholar] [CrossRef] [Green Version]
  46. Zhao, W.; Gao, B.; Liu, M.; Lu, Q.; Fan, S.J. Impact of Meteorological Factors on the Ozone Pollution in Hong Kong. Huan Jing Ke Xue = Huanjing Kexue 2019, 40, 55–66. [Google Scholar] [PubMed]
  47. Kang, S.M.; Polvani, L.M.; Fyfe, J.C.; Sigmond, M. Impact of Polar Ozone Depletion on Subtropical Precipitation. Science 2011, 332, 951–954. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  48. Juran, S.; Sigut, L.; Holub, P.; Fares, S.; Klem, K.; Grace, J.; Urban, O. Ozone Flux and Ozone Deposition in a Mountain Spruce Forest Are Modulated by Sky Conditions. Sci. Total Environ. 2019, 672, 296–304. [Google Scholar] [CrossRef]
  49. Bloomfield, P.; Royle, J.A.; Steinberg, L.J.; Yang, Q. Accounting for Meteorological Effects in Measuring Urban Ozone Levels and Trends. Atmos. Environ. 1996, 30, 3067–3077. [Google Scholar] [CrossRef]
Figure 1. Seven regions in China, namely, Central China (CC), East China (EC), North China (NC), South China (SC), Northwest China (NWC), Southwest China (SWC), and Northwest China (NEC); no data are available for Taiwan Province or Nanhai Zhudao.
Figure 1. Seven regions in China, namely, Central China (CC), East China (EC), North China (NC), South China (SC), Northwest China (NWC), Southwest China (SWC), and Northwest China (NEC); no data are available for Taiwan Province or Nanhai Zhudao.
Atmosphere 13 00908 g001
Figure 2. Spatial distributions of the annual average ozone concentration in 366 cities in China from 2016 to 2020 (red marks indicate that the concentration is higher than 160 μg·m−3, yellow marks indicate that the concentration ranges from 100 to 160 μg·m−3, and green marks indicate that the concentration is lower than 100 μg·m−3).
Figure 2. Spatial distributions of the annual average ozone concentration in 366 cities in China from 2016 to 2020 (red marks indicate that the concentration is higher than 160 μg·m−3, yellow marks indicate that the concentration ranges from 100 to 160 μg·m−3, and green marks indicate that the concentration is lower than 100 μg·m−3).
Atmosphere 13 00908 g002
Figure 3. Seasonal trends in the average ozone concentration (the y-axis represents the seasonal ozone concentration; units: μg·m−3).
Figure 3. Seasonal trends in the average ozone concentration (the y-axis represents the seasonal ozone concentration; units: μg·m−3).
Atmosphere 13 00908 g003
Figure 4. Monthly total number of cities in each region exceeding the O3 standard (160 μg·m−3) from 2016 to 2020 (unit: block).
Figure 4. Monthly total number of cities in each region exceeding the O3 standard (160 μg·m−3) from 2016 to 2020 (unit: block).
Atmosphere 13 00908 g004
Figure 5. Hourly concentration change trends of O3 and the other pollution factors (the left y-axis represents the O3 concentration; the right y-axis represents the concentrations of the other pollutants; units: μg·m−3).
Figure 5. Hourly concentration change trends of O3 and the other pollution factors (the left y-axis represents the O3 concentration; the right y-axis represents the concentrations of the other pollutants; units: μg·m−3).
Atmosphere 13 00908 g005
Figure 6. Histograms of the distributions of the correlation coefficients between the O3 concentration and various meteorological factors throughout the year and in each season for each of the seven regions of China.
Figure 6. Histograms of the distributions of the correlation coefficients between the O3 concentration and various meteorological factors throughout the year and in each season for each of the seven regions of China.
Atmosphere 13 00908 g006
Figure 7. Spatial clustering of annual and seasonal O3 concentrations in Chinese cities.
Figure 7. Spatial clustering of annual and seasonal O3 concentrations in Chinese cities.
Atmosphere 13 00908 g007
Table 1. Spatial autocorrelation indexes of the seasonal average O3 concentrations in 366 Chinese cities.
Table 1. Spatial autocorrelation indexes of the seasonal average O3 concentrations in 366 Chinese cities.
Value SpringSummerAutumnWinterYear
Moran’s I20160.170.230.400.170.24
20170.360.270.490.240.34
20180.300.280.690.210.37
20190.310.310.580.260.36
20200.370.300.680.180.38
Z[I]201617.9224.7145.3929.429.36
201732.7336.5457.2142.6642.29
201830.4035.1273.1337.2043.96
201934.7540.0069.1546.5547.61
202035.9937.5866.8729.6442.52
Table 2. Spatial autocorrelation indexes of the monthly average O3 concentrations in 366 Chinese cities.
Table 2. Spatial autocorrelation indexes of the monthly average O3 concentrations in 366 Chinese cities.
Value 123456789101112
Moran’s I20160.030.110.050.190.140.170.230.220.50.120.270.28
20170.250.250.140.360.320.270.210.210.380.550.380.30
20180.250.160.140.330.250.310.240.230.630.670.410.10
20190.270.040.210.190.270.300.280.240.530.730.510.30
20200.100.070.140.180.240.250.280.240.620.580.430.26
Z[I]20160.030.110.050.190.140.170.230.220.50.120.270.28
20170.250.250.140.360.320.270.210.210.380.550.380.30
20180.250.160.140.330.250.310.240.230.630.670.410.10
20190.270.040.210.190.270.300.280.240.530.730.510.30
20200.100.070.140.180.240.250.280.240.620.580.430.26
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Ge, Q.; Zhang, X.; Cai, K.; Liu, Y. Ozone Pollution in Chinese Cities: Spatiotemporal Variations and Their Relationships with Meteorological and Other Pollution Factors (2016–2020). Atmosphere 2022, 13, 908. https://doi.org/10.3390/atmos13060908

AMA Style

Ge Q, Zhang X, Cai K, Liu Y. Ozone Pollution in Chinese Cities: Spatiotemporal Variations and Their Relationships with Meteorological and Other Pollution Factors (2016–2020). Atmosphere. 2022; 13(6):908. https://doi.org/10.3390/atmos13060908

Chicago/Turabian Style

Ge, Qiang, Xusheng Zhang, Kun Cai, and Yang Liu. 2022. "Ozone Pollution in Chinese Cities: Spatiotemporal Variations and Their Relationships with Meteorological and Other Pollution Factors (2016–2020)" Atmosphere 13, no. 6: 908. https://doi.org/10.3390/atmos13060908

APA Style

Ge, Q., Zhang, X., Cai, K., & Liu, Y. (2022). Ozone Pollution in Chinese Cities: Spatiotemporal Variations and Their Relationships with Meteorological and Other Pollution Factors (2016–2020). Atmosphere, 13(6), 908. https://doi.org/10.3390/atmos13060908

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop