*3.2. Spatial Distribution Characteristics of O3*

#### 3.2.1. Spatial Aggregation Characteristics

To analyze the spatial aggregation and trends of O3, we performed a global spatial autocorrelation analysis of the average annual O3 concentrations. As a result, we obtained Moran's *I* and Z values. Table 2 demonstrates that Moran's *I* from 2016 to 2020 were 0.17, 0.36, 0.45, 0.45, and 0.51, respectively, while Z values were all >2.58 with an upward trend (*p* < 0.01). This indicates that the administrative units with high O3 concentrations or those with low O3 concentrations were significantly clustered in space (e.g., they are spatially positive). The O3 spatial aggregation increased annually in recent years. We noted a strong spatial autocorrelation and possible spatial aggregation.

**Table 2.** Spatial autocorrelation test results.


To further discuss the spatial aggregation of O3 concentrations, this study analyzed the cold hotspot map for O3 concentrations in the NCP; the results are shown in Figure 3. As shown in the figure, the hotspots in the NCP in 2016–2017 were mainly concentrated in the southwestern area of Shandong bordering Henan and Anhui Province. Jiangsu Province was also relatively concentrated. From 2018 to 2020, hotspots were concentrated in the western part of Shandong Province and the junction area of Henan Province and Hebei. From 2016–2020, cold spots were concentrated in Beijing, Tianjin, and surrounding areas, while the agglomeration in other areas was not prominent. Furthermore, after 2016, the spatial correlation of O3 in the western part of Shandong became increasingly more pronounced, and the aggregation characteristics were the most prominent in 2018, showing

the clustering characteristics of the Hebei–Henan–Shandong–Anhui region. Overall, the comprehensive 2016–2020 trends of O3 spatial aggregation changes show that the western area of Shandong bordering other provinces forms a stable high-value aggregation feature, whereas Beijing, Tianjin, and the surrounding areas form a stable low-value aggregation. These two parts of the regional spatial correlation of O3 concentrations were substantial and not easily influenced by other regions.

**Figure 3.** Spatial aggregation characteristics of O3 concentrations in the NCP from 2016–2020.

3.2.2. Annual Variation of the Spatial Distribution of O3

The spatial distribution of O3 in the NCP is shown in Figure 4. As shown, the spatial distribution of O3 was approximately the same and consistent with the spatial aggregation trend (Figure 3) from 2016 to 2020. Overall, the study area exhibited a spatial variation trend characterized by low O3 concentrations in the north and south and high concentrations in the central area. Among these, O3 concentrations in the west of the Shandong Province and the junction of the Jiangsu, Shandong, Hebei, Henan, and Anhui provinces were high, with a maximum value of 120.79 μg/m3. This area was characterized by high O3 concentrations, which are related to the presence of industrialized cities in the region. Meanwhile, the O3 concentrations in Beijing, southern cities of Henan Province, Anhui Province, Jiangsu Province, and coastal areas of Shandong Peninsula were low, with the lowest value being 85.58 μg/m3. Furthermore, O3 concentrations in the study area significantly increased from 2016 to 2018, while the high-value area of O3 concentrations gradually expanded. In 2019, O3 concentrations in Jiangsu Province and some surrounding cities in the south of the NCP decreased. In 2020, the pattern changed, and the overall O3 concentrations in the study area decreased. In previous years, O3 concentrations in Henan and Hebei

Provinces with higher O3 concentrations significantly decreased, while some areas with high concentrations persisted in the central and western regions of the Shandong Province.

**Figure 4.** Annual spatial distribution map of O3 in the NCP from 2016 to 2020: (**a**) Spatial distribution of O3 in 2016, (**b**) Spatial distribution of O3 in 2017, (**c**) Spatial distribution of O3 in 2018, (**d**) Spatial distribution of O3 in 2019, (**e**) Spatial distribution of O3 in 2020.

Precursors are essential for O3 formation. The sources of O3 precursors can be fundamentally divided into natural and anthropogenic sources. Natural sources include soil, lightning, and plant emissions, while anthropogenic sources include motor vehicle exhaust, coal combustion, and industrial and power plant emissions [35]. As the compilation of the VOC source emission inventory was delayed and underwent changes every year, no latest VOC source emission data were available for 2020. Due to this, the NOx emissions data were taken as the measurement index of precursors. Note that, as a precursor of O3, the NOx concentration is closely related to that of O3 [36,37]. The total NOx emissions comprise industrial, motor vehicle, and domestic emissions. According to data from the China Environmental Statistics Yearbook (2019), industrial and motor vehicle emissions account for more than 90% of the total emissions in China. Since municipal-level data on NOx emitted by motor vehicles were not available, industrial NOx emissions and civil vehicle ownership were selected as relevant indicators to measure the change in O3 concentrations. Given the lack of statistical data for some years, we considered the data of industrial NOx emissions in 2017 and civil vehicle ownership in 2019 as examples to shed light on the impact of O3 precursors on the spatial distribution of O3. The results are shown in Figure 5.

**Figure 5.** (**a**) Spatial distribution of civil vehicle ownership in the NCP in 2019; (**b**) Spatial distribution of industrial NOx emissions in the NCP in 2017.

Figure 5a shows the civilian car ownership map. As seen, the car ownership in Beijing– Tianjin–Hebei is generally high. The capital city of Beijing, being the political and cultural center of China, is characterized by the highest car ownership of 5.908 million, followed by Zhengzhou (Henan Province), with car ownership of 3.814 million. Jinan, Linyi and the Shandong Peninsula are also characterized by high car ownership. This pattern is driven by the population density of the Beijing–Tianjin–Hebei urban agglomeration. The area is large, and the degree of social and economic development is also high. As a highly populated province, the Shandong Province had a population of >100 million people in 2017 (Chinese Statistic Year, 2018), and motor vehicles were used extensively. As a result, these regions are characterized by the largest car ownership, whereas car ownership in most regions of the Henan and Anhui Provinces is low (the lowest car ownership is equal to 282,300 in the Hebi City of the Henan Province). The distribution map of industrial NOx emissions (Figure 5b) indicated that Tianjin, Tangshan, Qinhuangdao, southern Hebei Province, and central Shandong Province, such as Binzhou and Zibo, and other cities had higher industrial NOx emissions. Of these administrative units, Tangshan was characterized by the strongest annual emissions (196,572 tons/year), followed by Tianjin (73,249 tons/year). These cities are heavily industrialized with an economy focused on structure, metallurgy, chemical industry, building materials, and high-energy consuming industries, causing emissions of large amounts of pollutants. However, in the southern part of the study area (the Henan and Anhui Provinces), the emissions of industrial NOx were relatively low, with the lowest value being 2296 tons. Moreover, the emissions of industrial NOx in Beijing and the Shandong Peninsula were also relatively low. We combined the number of heavy industrial plants in each province in the study area, shown in Figure 6, and showed that Shandong Province had the largest number of heavy industrial plants, followed by Hebei Province. We argue that this is a manifestation of the convergence of the industrial structure between the provinces and the cities in Beijing–Tianjin–Hebei and the surrounding areas.

**Figure 6.** Number of heavy industrial plants in each province of the NCP.

Combined with the analysis results in Figure 4, the areas with high O3 concentrations were mainly clustered in the central and western regions of Shandong Province, the southern part of Hebei Province, and the junction of Jiangsu, Shandong, Henan, and Anhui Provinces. This finding is in line with the regions characterized by a large number of motor vehicles and large industrial NOx emissions. However, although Beijing and Tianjin were characterized by a large number of motor vehicles, the O3 concentrations in the area remained low. This pattern was driven by the strict motor vehicle emission reduction policies implemented by key cities such as Beijing and Tianjin, which effectively controlled the pollution of motor vehicle emissions [38]. Generally, precursor emissions are closely related to O3 concentrations [39]. Large-scale industrial production and massive traffic flow lead to larger NOx emissions. Namely, the greater the NOx emissions in the region, the more conducive the photochemical reaction conditions are and the greater the O3 concentration. However, the Qinhuangdao City (Hebei Province) was characterized by a large amount of industrial NOx emissions, and the Shandong Peninsula had a large number of motor vehicles. Thus, one could anticipate that more emissions of precursors could emerge in these areas, and O3 concentrations would inevitably increase. Nevertheless, O3 concentrations remained at a moderately low level. As these cities are close to the ocean and experience good atmospheric diffusion conditions, clean ocean air masses moderately exert a dilution effect on local pollution sources.
