Study on the Spatio-Temporal Characteristics and Driving Factors of PM_2.5 in the Inter-Provincial Border Region of Eastern China (Jiangsu, Anhui, Shandong, Henan) from 2022 to 2024

Abstract

The inter-provincial border region in eastern China, encompassing the junction of Jiangsu, Anhui, Shandong, and Henan provinces, serves as a crucial zone that connects the important economic zones of Beijing–Tianjin–Hebei and the Yangtze River Delta. It is of great significance to study the temporal variation characteristics, spatial distribution patterns, and driving factors of PM_2.5 concentrations in this region. Based on the PM_2.5 concentration observation data, ground meteorological data, environmental data, and socio-economic data from 2022 to 2024, this study conducted in-depth and systematic research by using advanced methods, such as spatial autocorrelation analysis and geographical detectors. The research results show that the concentration of PM_2.5 rose from 2022 to 2023, but decreased from 2023 to 2024. From the perspective of seasonal variations, the concentration of PM_2.5 shows a distinct characteristic of being “high in winter and low in summer”. The monthly variation shows a “U”-shaped distribution pattern. In terms of spatial changes, the PM_2.5 concentration in the inter-provincial border region of eastern China (Jiangsu, Anhui, Shandong, Henan) forms a gradient difference of “higher in the west and lower in the east”. The high-concentration agglomeration areas are mainly concentrated in the Henan part of the study region, while the low-concentration agglomeration areas are distributed in the eastern coastal parts of the study region. The analysis of the driving factors of the PM_2.5 concentration based on geographical detectors reveals that the average temperature is the main factor affecting the PM_2.5 concentration. The interaction among the factors contributing to the spatial differentiation of the PM_2.5 concentration is very obvious. Temperature and population density (q = 0.92), temperature and precipitation (q = 0.95), slope and precipitation (q = 0.97), as well as DEM and population density (q = 0.96), are the main combinations of factors that have continuously affected the spatial differentiation of the PM_2.5 concentration for many years. The research results from this study provide a scientific basis and decision support for the prevention, control, and governance of PM_2.5 pollution.

1. Introduction

Air pollution is a global environmental problem, threatening the human living environment and public health, due to an unprecedentedly severe situation [1,2,3]. Among air pollutants, PM_2.5 has become the main culprit endangering human health. This is due to its unique physical properties and chemical composition [4,5,6,7]. PM_2.5 can significantly reduce visibility, affect traffic safety, and cause inconvenience to human daily life. Meanwhile, as an important component of air pollutants, it also poses a serious threat to the balance and stability of the ecosystem [8,9]. Therefore, in-depth research on the current status of PM_2.5 pollution is particularly important. The junction area between Jiangsu, Anhui, Shandong, and Henan is a key area. It connects Beijing–Tianjin–Hebei, the Yangtze River Delta, and other important economic zones, making its geographic location very important [10,11]. This area encompasses 22 cities, located across four provinces (detailed in Section 2.1). These cities have shown a remarkable growth trend in terms of their economic development and population aggregation. These cities show strong economic growth. Population aggregation is increasing rapidly. However, industrialization and urbanization are also accelerating. Consequently, PM_2.5 pollution has become more prominent [12,13]. At present, the research on the distribution characteristics and influencing factors of PM_2.5 concentrations in this area is not clear enough and in-depth exploration is urgently needed.

In recent years, scholars have conducted a large number of studies on the spatio-temporal characteristics and influencing factors of PM_2.5 concentrations. Xu et al. [13] used trend analysis and the geographical detector method, combined with multi-source data, such as PM_2.5 remote sensing data, meteorological data, and DEM, to conduct a comprehensive evaluation of the spatio-temporal pattern changes and influencing factors of PM_2.5 concentrations in the Beijing–Tianjin–Hebei urban agglomeration from 2000 to 2021. Zhou et al. [14] explored the driving factors of PM_2.5 concentrations in 2015 by using the geographical detector method. The results show that the vegetation index, the number of buses, and electricity consumption are the main driving factors. Natural factors and their interaction with human activities play a decisive role in the changes to PM_2.5 concentrations in Guangzhou City. Li et al. [15] used geographical detectors in 283 cities in China to analyze the spatio-temporal differentiation of PM_2.5 concentrations and its influencing factors. The results indicate that fewer cities exceed the relevant standards. PM_2.5 shows an overall positive agglomeration. HH-type agglomeration occurs in Sichuan Basin, the North China Plain, and the middle and lower reaches of the Yangtze River, as well as the LL type in the northeast, southeast, and southwest. Geographical exploration indicates that the annual average temperature is the most decisive factor. Tai et al. [16], based on the PM_2.5 observation data on the United States from 1998 to 2008, applied a multiple linear regression model to explore the correlation between the PM_2.5 concentration and its components and meteorological variables. It was found that daily variations in meteorology can account for up to 50% of PM_2.5 variations, and key meteorological predictors were identified. Owoade et al. [17], based on the sample data on fine and coarse particulate matter collected in multiple locations in Nigeria from 2006 to 2013, used the positive matrix decomposition method to identify the sources of PM and determined that soil, biomass combustion, vehicle emissions, sea salt, and waste processing were the main sources. Deshmukh et al. [18], based on the concentration data on PM₁, PM_2.5, and PM_2.5–10 in ambient air in central India from July 2009 to June 2010, studied the seasonal distribution trends of particulate matter and their relationship with meteorological variables through correlation analysis, and found that the concentrations were higher in winter and were significantly affected by biomass combustion. Meanwhile, PM_2.5 is highly positively correlated with PM₁. Although existing studies have focused on the current situation in terms of PM_2.5 pollution in the inter-provincial border region of Jiangsu, Anhui, Shandong, and Henan (e.g., [19,20]), these studies are mostly limited to a single time period or qualitative analysis only. By integrating multi-source data and geographical detectors, this study systematically reveals the spatio-temporal variations in PM_2.5 concentrations and the interaction effects of its driving factors in the geographical area for the first time. The research results can provide a scientific basis for cross-regional collaborative governance.

In summary, significant progress has been made in the study of the spatio-temporal variations of PM_2.5 and its influencing factors in recent years. However, due to the differences in the geographical environment, climate conditions, and human activities, the causes of pollution in different regions are different, especially in regard to the analysis of specific regions, such as the situation in the inter-provincial border region of eastern China (Jiangsu, Anhui, Shandong, Henan), which is more complex. This region has achieved remarkable industrialization and urbanization. However, this progress has come with environmental costs [19]. Therefore, this study aims to: (1) quantify the spatio-temporal patterns in the PM_2.5 concentrations across the inter-provincial border region of Jiangsu, Anhui, Shandong, and Henan from 2022 to 2024; (2) identify the key driving factors influencing the spatial differentiation of PM_2.5 concentrations within this region, with a specific focus on the interactions between natural and socio-economic factors; and (3) apply advanced spatial analysis methods, particularly geographical detectors, to rigorously assess the individual and interactive effects of meteorological, topographical, environmental, and socio-economic variables on PM_2.5 levels.

2. Data and Methods

In this study, we aimed to clarify the spatio-temporal characteristics of PM_2.5 concentrations during 2022–2024 in the inter-provincial border region of eastern China. Furthermore, we systematically investigated the driving factors behind the dynamic variations in PM_2.5 concentrations. To this end, the results provide scientific support for regional air quality improvement and sustainable development strategies. Figure 1 presents the detailed research framework.

2.1. Study Area

The border area between Jiangsu, Anhui, Shandong, and Henan (32°10′ N–36 N, 114°20′ E–120°15′ E) is located in the transition area of the core economic belt in eastern China. Its spatial span starts from the Yellow River Delta in Dongying City, Shandong Province, in the north. It extends to the northern foot of the Dabie Mountains in Xinyang City, Henan Province, in the south. To the west, it reaches the southern side of the Funiu Mountains in Nanyang City, Henan Province. In the east, it extends to the Yellow Sea coast in Lianyungang, Jiangsu Province. It stretches approximately 1100 km from north to south and 900 km from east to west, with a total area of about 486,000 square kilometers. As shown in Figure 2, the study area covers 22 cities in 4 provinces, including Xuzhou, Lianyungang, and Suqian in Jiangsu Province; Huaibei, Fuyang, Suzhou, and Bozhou in Anhui Province; Qingdao, Zaozhuang, Dongying, Weifang, Tai’an, Linyi, and Rizhao in Shandong Province; and Pingdingshan, Xuchang, Luohe, Nanyang, Shangqiu, Xinyang, Zhoukou, and Zhumadian in Henan Province [20,21].

The study area belongs to the warm temperate semi-humid monsoon climate. The annual average temperature is generally between 14 °C and 17 °C. The temperature is relatively low in winter and relatively high in summer. Annual precipitation in the border areas ranges from 600 mm to 1200 mm. It shows a distribution pattern of higher precipitation in the south and less in the north. Similarly, precipitation is higher in the east and lower in the west. The precipitation is mainly concentrated from June to September (accounting for more than 70% of the annual total). From the perspective of industrial structure, the industrial structure of the cities in this region is relatively highly concentrated. Among them, half of the cities have steel enterprises, half have coking enterprises, and two-thirds have cement clinker production enterprises. Overall, the cities in the border areas of Jiangsu, Anhui, Shandong, and Henan provinces have large scale industrial activities and relatively high pollutant emissions.

2.2. Research Data

This study adopted the daily PM_2.5 concentration data provided by the National Environmental Monitoring Center from 2022 to 2024, covering 22 cities in the border area of Jiangsu, Anhui, Shandong, and Henan provinces.

The detection factors in this study mainly include DEM, slope, aspect, temperature, precipitation, vegetation index, gross national product, and population density. The DEM and NDVI data were derived from the Cloud Platform for Resources and Environment Data at the Chinese Academy of Sciences (https://www.resdc.cn (accessed on 15 October 2024)). Its spatial resolution is 1 km, and the slope and aspect are extracted using ArcGIS 10.8 software, based on DEM data, as shown in Figure 3. The meteorological data (temperature, precipitation), gross national product, and population density are derived from the statistical yearbooks for each province.

2.3. Research Method

2.3.1. Spatial Autocorrelation Analysis

Spatial autocorrelation indicates that adjacent observation units have similar variable values [22,23,24,25,26]. It is often used to reveal the spatial dependence and spatial heterogeneity of regional variables. The calculation formula for the global Moran’s I index is [27]:

$$I={\displaystyle \frac{n\sum _{i=1}^n\sum _{j=1}^n{w}_{ij}(x_i-\overline{x})(x_j-\overline{x})}{\sum _{i=1}^n\sum _{j=1}^n{w}_{ij}\sum _{i=1}^n(x_i-\overline{x}{)}^2}}$$

(1)

In the formula, x_i and x_j represent the PM_2.5 concentrations in city i and city j, respectively; x is the average value of the PM_2.5 concentration in all the cities; n is the number of research units; and w_ij is the spatial weight matrix. The value of I ranges from −1 to 1; I > 0 indicates positive autocorrelation, I < 0 indicates negative autocorrelation, and I close to 0 indicates random distribution.

The local Moran’s I index corresponds to four types of associations in the quadrants of the scatter plot: high–high (HH) agglomeration, low–low (LL) agglomeration, low–high (LH) agglomeration, and high–low (HL) agglomeration. Among them, LL (HH) indicates that there is a spatial agglomeration effect in cities with low (high) PM_2.5 pollution, and LH (HL) indicates that cities with low (high) PM_2.5 pollution are surrounded by cities with high (low) pollution.

Hot spot analysis, based on Getis-Ord Gi*, is another method [28] used to test the existence of spatial effects. It identifies the hot spots and cold spots where PM_2.5 pollution accumulates, from a local perspective.

2.3.2. Geographical Detector

A geographical detector is a model that detects the spatial differentiation influence of independent variables on geographical phenomena [29,30]. The basic principle is that if the independent variable has a strong influence on the dependent variable, then there will also be a certain similarity in their spatial distribution. This study utilized factor detection and interaction detection in regard to the geographical detector to quantitatively investigate the influence of each influencing factor and the interaction between factors on the spatial differentiation of PM_2.5 in the border area of Jiangsu, Anhui, Shandong, and Henan in 2023. The calculation formula for factor detection is:

$$q=1-{\displaystyle \frac{{\sum }_{h=1}^L{N}_h\sigma \genfrac{}{}{0pt}{}{2}{h}}{N{\sigma }^2}}$$

(2)

In the formula, q represents the influence of the factor on the spatial differentiation of PM_2.5, with a value range of 0–1. The driving force intensity was quantified by the q-value in regard to the geographical detector, where higher values indicate a stronger influence. The larger the q value, the greater the influence of this factor on the spatial differentiation of PM_2.5. The geographical contribution rate is the percentage of PM_2.5 variance explained by each factor. Where h is the number of sub-regions of the detection factor X; L represents the stratification of the variables; N and N_h are, respectively, the total sample size and the sample size of region h; and σ² and σ_h² are, respectively, the variance of the total region and the variance of region h.

The interaction module in the geographical detector is mainly used to study the magnitude of the influence of the synergy between the factors on independent variables. The natural factors include DEM (X1), slope (X2), aspect (X3), temperature (X4), precipitation (X5), and the vegetation index (X6). The human factors include gross national product (X7) and population density (X8).

3. Results and Discussion

3.1. Temporal Variation Characteristics of PM_2.5

The interannual variation in the PM_2.5 concentration from 2022 to 2024 is shown in Figure 4. The concentration of PM_2.5 rose from 2022 to 2023 and decreased from 2023 to 2024. The PM_2.5 concentration values in 2022, 2023, and 2024 were 40.3 μg/m³, 41.8 μg/m³, and 40.7 μg/m³, respectively. To decouple the impacts of anthropogenic activities and meteorological conditions on PM_2.5 variations, we analyzed auxiliary data from the China Meteorological Administration (CMA) during 2022–2024. Wind speed/direction: The low wind speeds (<1.5 m/s) that occurred in 2023 (23% of days), compared to 2024 (15%) and 2022 (22%), potentially enhanced local pollutant accumulation. The annual precipitation in 2023 was 5% lower than in 2024, reducing the wet deposition of aerosols. The observed PM_2.5 increase in 2023 was strongly driven by the industrial rebound that occurred during the post-pandemic recovery. The key emission surges were documented: NOx emissions rose by 17% in Jiangsu and 9% in Henan, while particulate emissions spiked by 31% in Henan. This industrial expansion aligns with the findings by Shikhovtsev et al. [31] on production resumption effects. The synergy between elevated emissions and unfavorable meteorology (low winds, reduced precipitation) created compound pollution conditions, as demonstrated in the work by Wei et al. [32].

The changes in the PM_2.5 concentrations in 22 cities within the study area from 2022 to 2024 are shown in Figure 5. At the city level, significant differences in PM_2.5 concentrations were observed across the study area. Coastal cities generally experienced lower PM_2.5 concentrations compared to inland cities. For instance, the PM_2.5 concentrations in cities like Qingdao, Rizhao, and Lianyungang dropped from 20 to 30 μg/m³, while those in Zhumadian, Shangqiu, and Xuchang rose against the trend from 44 to 50 μg/m³ in 2024.

The seasonal variation in the PM_2.5 concentration values in the study area from 2022 to 2024 is shown in Figure 6. The concentration value of PM_2.5 shows a distinct seasonal variation, being higher in winter and lower in summer, throughout the year. This change may be related to seasonal meteorological conditions, industrial emissions, traffic conditions, and activities such as heating and burning. In winter, due to the increased demand for heating and the increased combustion activities, the concentration value of PM_2.5 may rise. In summer, the temperature is high and the atmospheric diffusion conditions are good, which is conducive to the diffusion of pollutants. Therefore, the concentration value is relatively low.

The monthly variation characteristics of the PM_2.5 concentrations from 2022 to 2024 are shown in Figure 7. Overall, the concentration of PM_2.5 was significantly higher from December to February of the following year, which was mainly related to the increase in air pollutant emissions caused by winter heating. From June to August, due to high temperatures and enhanced air circulation, the concentration of PM_2.5 reached the lowest level of the year. Unlike high-latitude or arid regions (such as those studied by Shikhovtsev et al. [31]; Wei et al. [32], the PM_2.5 concentration in the study area presented a ‘U-shaped’ seasonal distribution, and no secondary peak occurred in summer. This is mainly attributed to the low forest coverage rate (<10%) and very little fire activity in the region. Meanwhile, the situation was also influenced by heavy summer precipitation (accounting for 70% of the annual precipitation) and the clearance effect of monsoon circulation on pollutants. In addition, there are seasonal differences in regard to industrial emission sources (winter heating exacerbates emissions).

The number of PM_2.5 pollution days (PM_2.5 > 75 μg/m³) in each month from 2022 to 2024 is shown in Figure 8. The most polluted days occurred from November to February of the following year, accounting for 79.0%. The number of polluted days was the fewest from June to August, with only one day occurring in both June and August. To further characterize the monthly variations in the PM_2.5 concentrations, statistical summaries including the maximum, minimum, median, and mode values for each month from 2022 to 2024 were calculated and are illustrated in Figure 9. The results show that the PM_2.5 concentrations exhibited significant seasonal variation. The highest monthly median concentrations were observed in December and January, exceeding 70 μg/m³, while the lowest values occurred in June, July, and August, which were typically below 25 μg/m³. The maximum PM_2.5 concentrations were mostly recorded in the winter months, with some values surpassing 200 μg/m³, indicating severe pollution episodes. In contrast, the summer months showed not only lower medians, but also narrower interquartile ranges, suggesting more stable and cleaner air conditions. The mode values were generally consistent with the median trends, reinforcing the reliability of the seasonal patterns.

3.2. Spatial Variation Characteristics of PM_2.5

The spatial distribution of the PM_2.5 concentrations in the study area from 2022 to 2024 is shown in Figure 10. From 2022 to 2024, the PM_2.5 concentration in the border areas of Jiangsu, Anhui, Shandong, and Henan provinces was generally lower in the northeastern part of the region. Among them, the PM_2.5 concentration in Dongying was 34.4 μg/m³, Weifang was 35.7 μg/m³, Qingdao was 27.7 μg/m³, Rizhao was 30.4 μg/m³, and Lianyungang was 31.7 μg/m³. The PM_2.5 concentration was relatively high in the northwest region. Among them, the PM_2.5 concentration was 47.8 μg/m³ in Shangqiu, 46.6 μg/m³ in Zhoukou, 47.3 μg/m³ in Pingdingshan, and 48.1 μg/m³ in Xuchang.

The concentration of PM_2.5 in cities in the northwest was relatively high in 2022. Among them, the annual average concentration of PM_2.5 in Luohe, Pingdingshan, Xuchang, Shangqiu, and Nanyang in Henan Province was above 46.5 μg/m³. As shown in Figure 10c, the annual average regional PM_2.5 concentration rose overall in 2023. As shown in Figure 10d, the high values of the regional PM_2.5 concentration in 2024 were mainly located in the northwest, ranging from 44.3 to 49.2 μg/m³. Compared with 2022, the PM_2.5 concentration in Xinyang, Bozhou, and Nanyang decreased significantly, by 5.8%, 4.6%, and 4.5%, respectively, while the PM_2.5 concentration in Huaibei decreased the least, at 0.6%. From the perspective of the interannual spatial variation, the PM_2.5 concentration in cities in the northwest of the region decreased significantly.

The spatial distribution of PM_2.5 pollution days from 2022 to 2024 is shown in Figure 11. The number of polluted days in the southern inland cities was relatively high. For instance, Nanyang City (351 days), Luohe City (347 days), and Shangqiu City (343 days) all exceeded 340 pollution days, making them severely polluted areas. The pollution level in the eastern coastal cities has been significantly alleviated. In Qingdao City (120 days), Rizhao City (150 days), and Dongying City (168 days), the number of polluted days was less than 200.

The concentration of industrial activities explains the pollution in the southern region. Fossil fuel-dependent energy structures worsen air quality. Heavy industry and traffic emissions compound these effects. In addition to this, inland cities are restricted by the geographical conditions (such as frequent calm and windy days and poor diffusion conditions), which intensifies the local accumulation of pollutants. Meanwhile, the eastern coastal areas have achieved effective pollution control through the strengthening of environmental protection policies and the advantages of the natural conditions in the region. Furthermore, the infrastructure expansion and population concentration brought about by rapid urbanization may further increase the pollution control pressure on cities in the central and southern regions.

The seasonal spatial distribution of PM_2.5 concentrations in the border areas of Jiangsu, Anhui, Shandong, and Henan from 2022 to 2024 is shown in Figure 12. The concentration of PM_2.5 in this area is higher in spring and winter. Among them, the monthly average PM_2.5 concentration in winter was generally higher than 69.5 μg/m³. The high values in winter mainly result from the sharp increase in coal-burning pollution emissions during the heating period. Coupled with the reduction in the height of the boundary layer, the vertical diffusion of pollutants is restricted in winter and, thus, they accumulate. The concentration of PM_2.5 rose significantly in spring, with values ranging from 28.8 to 49.1 μg/m³. This increase is mainly due to cross-border dust transportation driven by Mongolian cyclones. In summer, affected by the strengthening of the southeast monsoon and the scouring effect of precipitation, the concentration of PM_2.5 dropped to the lowest level seen throughout the year.

The spatial agglomeration distribution of PM_2.5 concentrations in the border areas of Jiangsu, Anhui, Shandong, and Henan from 2022 to 2024 is shown in Figure 13.

As shown in the distribution map of high and low clusters in Figure 13a, there are a total of four cities with HH-type clusters, accounting for 18.2% of the total, mainly concentrated in Pingdingshan, Xuchang, Luohe, Zhoukou, and other places in Henan Province, at the junction of Jiangsu, Anhui, Shandong, and Henan. These regions are in the process of rapid industrialization. Resource-intensive industries dominate, resulting in large emissions of air pollutants and the formation of persistent high-pollution areas. There is a total of four cities in the LL-type cluster, accounting for 18.2% of the total. They are mainly distributed in the eastern coastal areas, such as Lianyungang, Rizhao, Weifang, and Qingdao. These areas belong to the eastern monsoon region, with relatively low pollution emissions and excellent air quality, forming stable low-value clusters.

In this study, hot spots indicate areas with significantly higher PM_2.5 concentrations (Getis-Ord Gi*, p < 0.05), while cold spots denote areas with significantly lower concentrations. As shown in the cold and hot spot distribution map in Figure 13b, the hot spot areas and cold spot areas of PM_2.5 are located in the northwest and northeast regions of this area, respectively. From the perspective of agglomeration trends, during the period from 2022 to 2024, the hot areas were mainly concentrated in the Henan urban agglomeration, while the cold spots were mainly distributed in the eastern coastal areas.

The data in

Table 1. Moran’s I index for each month in the study area from 2022 to 2024.

Month	1	2	3	4	5	6	7	8	9	10	11	12
Moran’s I	0.79	0.65	0.37	0.41	0.46	0.38	0.34	0.57	0.62	0.83	0.69	0.45
Z (I)	3.20	2.70	1.66	1.82	1.98	1.66	1.51	2.37	2.63	3.39	2.85	1.92

show that Moran’s I index for all the months is positive (ranging from 0.34 to 0.83). It indicates that the PM_2.5 concentration shows a significant positive spatial correlation in this area, that is, high values or low values tend to be aggregated and distributed together. All the Z values in the table are greater than 1.5. Among them, the Z values in January, February, August, September, October, and November exceed 2.0, indicating that the spatial autocorrelation of these months is statistically significant.

The seasonal trends in Moran’s I index are higher in cold months, which includes October, January, November, and February. This indicates stronger PM_2.5 spatial aggregation during colder periods. Moran’s I index is relatively low in summer (June and July), which may be related to meteorological conditions such as precipitation and diffusion.

3.3. Multi-Dimensional Detection of PM_2.5 Influencing Factors

3.3.1. Detection of PM_2.5 Impact Factors

Factor detector is utilized to determine the degree of influence of each influencing factor on the change in the PM_2.5 concentration. The natural factors include DEM (X1), slope (X2), aspect (X3), temperature (X4), precipitation (X5), and the vegetation index (X6). The human factors include gross national product (GDP) (X7) and population density (X8). The statistical results for the factor detection q values in the border area of Jiangsu, Anhui, Shandong, and Henan in 2023 are shown in

Table 2. Statistics on the factor detection q values in the study area in 2023.

Factor	DEM	Slope	Aspect	Temperature	Precipitation	Vegetation Index	GDP	Population Density
q	0.18	0.17	0.40	0.78	0.45	0.40	0.43	0.12

. The q value range of all the influencing factors is from 0.12 to 0.78. The order of their degree of influence is as follows: temperature (0.78) > precipitation (0.45) > gross national product (0.43) > vegetation index (0.40) > aspect (0.40) > topography (0.18) > slope (0.17) > population density (0.12). The influence of temperature on the PM_2.5 concentration is the most significant. In areas with suitable climates, higher temperatures usually enhance the vertical diffusion capacity of the atmosphere, which is conducive to the dilution of pollutants. In extremely low-temperature areas, stable inversion layers are prone to form, resulting in the continuous accumulation of PM_2.5 at the near-surface level. Coal heating in winter will directly emit a large amount of PM_2.5. Precipitation (q = 0.45) and the vegetation index (q = 0.40) jointly reflect the purification capacity of the natural environment. Precipitation removes particulate matter through wet deposition, while vegetation reduces the concentration of pollutants through adsorption and inhibition [33,34,35,36]. The gross national product (q = 0.43) indicates that the level of economic development is significantly correlated with the spatial distribution of pollution, and attention needs to be paid to the control of emissions in intensive industrial areas. The slope direction (q = 0.4) may affect the diffusion of pollutants by changing the local airflow, while the slope and population density have a relatively weak effect on pollutants (q ≤ 0.17).

The q value represents the explanatory power of each factor on the spatial variation in the PM_2.5 concentration. A higher q value indicates a stronger influence. In this study, temperature (q = 0.78) shows the highest explanatory power, suggesting that climatic conditions, especially temperature, play a dominant role in shaping the spatial pattern of PM_2.5. This reflects the transitional climate characteristics of the study area. Frequent inversion layers in winter (such as an average daily inversion intensity of 1.5 °C/100 m in Xuzhou in January 2023) have led to the continuous accumulation of PM_2.5. Although high temperatures in summer enhance diffusion, the intensification of the photolysis of ozone precursors during the same period may indirectly increase the generation of secondary aerosols. Precipitation (q = 0.45) shows moderate independent explanatory power. This figure is approximately 50% higher than that of the Yangtze River Delta region (q ≈ 0.3), which may be related to the fact that this area relies more on precipitation to remove air pollutants. In contrast, population density (q = 0.12) shows the weakest influence, possibly due to its indirect and spatially heterogeneous impact on PM_2.5, which may be masked by other dominant factors, such as GDP or temperature.

3.3.2. Interactive Detection of PM_2.5

The interactive detector detected the influence of the interaction of each influencing factor on the spatial distribution of the PM_2.5 concentration, and the results are shown in

Table 3. The average q value of the impact factor interaction in 2023.

Factor	DEM	Slope	Aspect	Temperature	Precipitation	Vegetation Index	GDP	Population Density
DEM	0.18
Slope	0.68	0.17
Aspect	0.86	0.69	0.40
Temperature	0.95	0.94	0.92	0.78
Precipitation	0.78	0.63	0.97	0.95	0.45
Vegetation index	0.77	0.66	0.62	0.83	0.67	0.40
GDP	0.75	0.73	0.85	0.86	0.71	0.67	0.43
Population density	0.96	0.86	0.86	0.92	0.67	0.61	0.65	0.12

. The interaction effect of any two influencing factors on the change in the PM_2.5 concentration is greater than the independent effect of a single factor. The interaction between natural factors and artificial factors is particularly prominent. The combinations of temperature and population density (q = 0.92), temperature and precipitation (q = 0.95), slope and precipitation (q = 0.97), as well as DEM and population density (q = 0.96), showed the strongest interaction effect, and their q values were all close to or exceeded 0.9. These findings indicate that the explanatory power of these factors in regard to the PM_2.5 concentration is significantly enhanced when these factors act together. In particular, the interaction between the slope aspect and precipitation shows typical nonlinear enhancement characteristics (q = 0.97 is greater than the sum of the single factors 0.85), indicating that in terrain-specific areas, the reduction in precipitation may intensify pollution accumulation. Furthermore, combinations of factors, such as temperature and vegetation index (q = 0.83), gross national product and slope aspect (q = 0.85), also show a two-factor enhancement effect, suggesting that vegetation coverage and economic activities have a superimposed impact on pollution distribution under specific environmental conditions. In contrast, the interaction between pure natural terrain factors (such as slope and DEM) is relatively weak (q = 0.68), indicating a limited joint contribution to the spatial differentiation of PM_2.5.

The interaction detector results reveal that the combined effects of two factors are always greater than their individual effects, especially when natural and anthropogenic factors interact. For example, the interaction between temperature and precipitation (q = 0.95) suggests that when subject to a high temperature and sufficient rainfall, the atmosphere’s self-purification capacity is significantly enhanced. The influence of temperature and precipitation on the PM_2.5 concentration is interrelated. Precipitation effectively removes particulate matter from the atmosphere through the wet deposition process, while temperature regulates the distribution of PM_2.5 by influencing the convective and diffusive capabilities of the atmosphere. Under high-temperature conditions, the wet deposition efficiency of precipitation is higher. Due to the enhanced thermal convection in these conditions, it helps to diffuse pollutants near the ground upwards, thereby reducing the concentration of PM_2.5. On the contrary, in low-temperature conditions, although precipitation can still reduce the PM_2.5 concentration through washing, the occurrence of a stable inversion layer may limit the vertical diffusion of pollutants, leading to the accumulation of PM_2.5 in the near-surface layer. This synergistic effect of temperature and precipitation (q = 0.95) indicates that the two have a significant combined influence on the spatial distribution of the PM_2.5 concentration, especially under the combined effect of seasonal changes and regional climatic conditions. Similarly, the strong interaction between the slope aspect and precipitation (q = 0.97) indicates that in areas with specific topographic features, reduced precipitation may exacerbate pollution accumulation, due to limited pollutant dispersion. These findings highlight the importance of considering factor interactions in regional pollution control strategies. The high interaction value (q = 0.92) of temperature and population density indicates that in densely populated areas that are using winter heating, such as Zhoukou and Fuyang, emission reduction efforts need to be combined with thermal regulation. The interaction effect of GDP+ aspect (q = 0.85) reveals that industrial cities (such as Xuzhou) need to avoid the park located in unfavorable diffusion terrain. The interaction between slope direction and precipitation shows a nonlinear enhancement (q = 0.97 > single factor and 0.85). This mainly stems from the regulation of wet settlement efficiency by the terrain. When the prevailing wind direction (such as the northwest wind in winter) is consistent with the sunny slope, the efficiency of precipitation in removing pollutants from the leeward slope decreases.

These results have important implications for regional pollution control. In areas with high temperatures and dense populations, priority should be given to implementing emission control, while in areas with little rainfall and unfavorable slopes, the monitoring of and interventions in regard to pollution diffusion conditions should be strengthened. Meanwhile, regions with high economic growth need to combine vegetation restoration measures to alleviate the synergistic pollution effect of human activities and natural factors.

4. Conclusions and Prospects

4.1. Conclusions

This study analyzed the spatio-temporal characteristics and driving factors of PM_2.5 concentrations in the border area of Jiangsu, Anhui, Shandong, and Henan from 2022 to 2024. The results show that the PM_2.5 concentrations exhibited a “high in winter, low in summer” seasonal pattern and a “U-shaped” monthly trend. Spatially, the PM_2.5 concentrations were higher in the west and lower in the east, with high-concentration clusters mainly located in the Henan urban agglomeration. The geographic detector analysis revealed that temperature was the most significant factor influencing the PM_2.5 concentration, and the interaction between natural and anthropogenic factors significantly enhanced the explanatory power of such factors.

These findings provide a scientific basis for regional air quality management and highlight the importance of cross-provincial coordination in regard to pollution control. However, this study has several limitations. First, the analysis is based on annual and seasonal averages, which may mask short-term pollution events. Second, the spatial resolution of some of the socio-economic data is limited to the city level, which may affect the accuracy of local-scale factor analysis. Third, the study period is relatively short (2022–2024), which may not fully capture long-term trends or interannual variability.

In future research, we will introduce machine learning algorithms (such as random forests, gradient boosting trees, etc.) for feature importance analysis. The aim is to identify more accurately the factors that have the greatest influence on PM_2.5 emission reduction. Then, the robustness of the relevant causal relationships will be further verified. Secondly, future research can involve the construction of spatial econometric models (such as spatial lag models and spatial error models). The aim is to more accurately capture the cross-regional spillover effects of policies and provide theoretical support for regional environmental governance cooperation. In order to break through the resolution limitation of the current county-level data, future research can integrate multi-source data through the use of the real-time monitoring network of the Internet of things.

4.2. Policy Recommendations

(1)

Strengthen the control of industrial emissions. Especially in intensive industrial areas (such as cities like Nanyang, Luohe, and Shangqiu in Henan Province), industrial emission standards should be raised further and pollutant emissions from highly polluting industries should be strictly restricted.

(2)

Strengthen vegetation restoration and ecological protection. In industrial cities and areas with serious pollution, urban green spaces and vegetation cover should be increased to improve the natural purification capacity of the environment. At the same time, efforts should be made to strengthen the protection of natural ecosystems, such as forests and wetlands, and give full play to their important role in purifying the air.

(3)

Improve the regional collaborative governance mechanism. Establish a regional air quality monitoring data-sharing platform to share monitoring data in real time, and improve the scientific and timeliness of regional collaborative governance. Promote cross-regional environmental governance projects, such as jointly launching air pollution prevention and control actions to jointly address regional air pollution issues.