Combined pollution of heavy metals and polycyclic aromatic hydrocarbons in the soil in Shenfu Region, China: a case of three different cities

It is a challenging issue to investigate the combined pollution of heavy metals and polycyclic aromatic hydrocarbons (PAHs) in urban soils. The purpose of this study was to determine the concentrations of these two pollutants in soils in Shenyang, Fushun, and Fushun New District, to analyze their distribution, their interaction, and co-contamination levels. The concentrations of heavy metals were measured by inductively coupled plasma mass spectrometry (ICP-MS), while the concentrations of 21 kinds of PAH were analyzed by gas chromatography-mass spectrometry (GC–MS). Based on the analysis of pollution concentrations and distribution patterns, the intrinsic links between heavy metals and PAHs in three different cities were assessed using a variety of multivariate analysis methods. Compared to Shenfu New District, the concentration of pollutants in Shenyang and Fushun shows a higher level. Moreover, the results of redundancy analysis (RDA) of samples may quantify the possibility of combined pollution of different heavy metal elements and PAHs. This study also affirms the important role of multivariate analysis in being used to reveal the complex interactions and spatial distribution of different pollutants.


Introduction
Soil, as an important carrier of pollutants in the environment, contains 90% of contaminants generated by mankind (Gutiérrez-Ginés et al., 2014). In the soil system, heavy metals and polycyclic aromatic hydrocarbons (PAHs) are two classes of contaminants with different properties . Both of them are trace and persistent pollutants in the environment, which are stable in the soil and easily adsorbed in soil organic matter (Yuesuo et al., 2017). In addition, these two groups of pollutants are also of wide international concern because of their health risks and ecotoxicity (Cao et al., 2017;Chen et al., 2013;Farago, 1994;Yongming et al., 2006;Zhu et al., 2016). Traditionally, most studies reported the concentrations of these two groups of pollutants separately . In fact, due to their similar chemical properties and sources, more new studies in recent years have focused on the coexistence of these two types of pollutants in the environment (Gulan et al., 2017;Thavamani et al., 2012;Wang et al., 2018).
Combined pollution refers to the coordinated pollution of multiple pollutants to the same medium (soil, water, atmosphere, organism) (Gutiérrez-Ginés et al., 2014), including metal-organic co-contamination. The existence of similar physicochemical similarities as well as homology between heavy metals and PAHs makes them susceptible to co-adsorption by soil colloids and thus accumulation in the soil. Many organisms are threatened by the synergistic cytotoxic effects of PAHs and heavy metals (Staninska-Pięta et al., 2020). The synergistic cytotoxic effects of PAHs and heavy metals have been shown to disrupt microbial cell membranes, enzyme specificity, metabolic pathway function, and retard protein and genetic activity, resulting in a significant reduction in soil microbial populations (Huang et al., 2019;Khatoon & Malik, 2019;Kim et al., 2019). A growing body of literature documents the uptake of these contaminants in soil by organisms through ingestion, respiration, and dermal uptake (Marinho Reis et al., 2016;Qu et al., 2020;Stajic et al., 2016). Understanding the retention and co-contamination of these two contaminants in soil is particularly important for human health and safety. Current studies on these two pollutants have mainly focused on their spatial distribution and ecological risks in some areas, while fewer types of studies have been conducted on the interactions between the two pollutants in specific regions. Wang Chen investigated the concentrations of heavy metals and PAHs in the soil around a cement plant in Beijing and used a risk assessment model to show the potential risk of pollutants to the population and environment in the vicinity of the cement plant (Wang et al., 2018). Similarly, in the Galicia region, northwest Spain, Monaco studied the PAHs, heavy hydrocarbons, and metal pollution levels in soils, and the contamination levels were shown to be high at several locations in this region (Monaco et al., 2017). Thavamani used multivariate analysis to examine the mixed contaminants (PAHs and heavy metals) at a manufactured gas plant site soils (Thavamani et al., 2012).
Along with China's economic growth, its urban construction has also grown rapidly in the past two decades. The aggregation of population and development in industry and sport caused by the booming of urbanization may change the sources and distribution patterns of soil pollutants (Cao et al., 2017). For example, Peng Chi found different distribution patterns of PAHs and heavy metals in urban soils during urbanization . Mahdi Ahmed also found the change of heavy metals and PAH baselines in sediment caused by urbanization (Mahdi Ahmed et al., 2017). Moreover, Hussain made a comprehensive assessment of PAHs, carbon, and metals in street dust, where vehicular emission, coal, and wood-burning were the main sources (Hussain et al., 2015). However, these reports treated the two groups of pollutants separately, and none of them mentioned the interaction of different pollutants. Therefore, it is important to study the co-pollution of PAHs and heavy metals in rapidly urbanizing areas.
The objectives of this study were (1) to discuss the contaminations of heavy metals and PAHs in the area; (2) to determine the distribution patterns of the two groups of pollutants; and (3) to analyze the cocontaminations and interactions between the two pollutants, which could provide a theoretical basis and data supporting for combined pollution research in the future work.

Study area and sampling
Shenyang and Fushun are the most important industrial cities in Northeast China. The total population of the two cities exceeds 10 million. Since 1950, after a long period of industrialization and urbanization, the region has formed several industries, such as petrochemical, pharmaceutical, metallurgical and electroplating, printing, and dyeing industries. Studies have shown that each of these industries is a significant source of heavy metals and PAHs (Liu et al., 2015).
Soil samples (Fig. 1) were collected from Shenyang, Fushun, and Shenfu New District, with a total area of 6079 km 2 . Shenfu New District is located in the middle region of Shenyang and Fushun, which is a satellite city of Shenyang under rapid urbanization (Sang et al., 2009). The urbanization rate of the Shenandoah New Area increased from 5% to over 40% from 2010 to 2020. Considering the impact of the land use type, transportation, and industrial distribution, we choose 66 sampling sites based on a 0.01° × 0.01° grid constructed by ArcGIS 10.2. A total of 66 topsoil samples (0~20 cm) were collected from different sites in mid-November 2016 (3 background spots samples, 29 Shenfu New District samples, 15 Shenyang samples, and 19 Fushun samples). The stones, plant roots, and other large debris were removed and the soils were put into aluminum boxes and plastic bags. All the samples were transferred to the laboratory and stored at − 20 °C until analysis.

Chemical analysis of heavy metals and PAHs
The analytical method of heavy metals was adopted from a previous study (Zang et al., 2016). All the samples were freeze-dried and sieved through a 100 mesh before analysis. Approximately 0.10 g of dried soils was ground and transferred into clean and dry digestion tubes. After being added with concentrated HNO 3 (8 mL) and stabilized overnight, the tubes were then placed on a heating block (Mars-Xpress, CEM) for digestion. The digestion temperature program were shown in Table S1. After digestion, the solutions were cooled, diluted to 50 mL with ultrapure water, and finally filtered into plastic bottles prewashed with nitric acid. The concentrations of heavy metals (Cr, Ni, Cu, Zn, As, and Pb) in the acid digests were measured by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7500a).
The extraction and clean-up of PAHs were also adopted from a previous study (Ma et al., 2009). Briefly, all the samples were sieved through a 100 mesh. Ten grams of samples was mixed with anhydrous sodium sulfate and spiked with 100 μL of the labeled surrogated standard mixture. All the samples were then Soxhlet extracted for 24 h with a 200 mL mix of solvent (n-hexane/acetone, 1:1, v/v). The extracts were concentrated and cleaned up by a selfpacked column, containing, from top to bottom, 10 g of activated silica gel and 2 g of anhydrous sodium sulfate. The extract was eluted with a 60 mL mixture of dichloromethane and hexane (1:1, v/v). The elution was finally concentrated to 1 mL under nitrogen evaporation before gas chromatography-mass spectrometry (GC-MS) analysis.

Quality control
For quality control of heavy metals, analytical blank and standard reference soil (from the Chinese National Standard Soil Bank, GBW07402 GSS-2 and GSS3) were performed in each sample batch to validate the digestion procedure and metal analysis. Recoveries were 83.1~111.6% for Cr, 73.6~140.1% for Ni, 94.9~128.1% for Cu, 95.3~139.8% for Zn, 75.4~122.5% for As, and 68.2~124.7% for Pb. The results showed that the detected values with the accuracy requirements compared to the certified values.
Quantification of PAHs was performed by a fivepoint calibration curve (R 2 > 0.999). Strict quality control procedures were conducted in this study. For every batch of 10 samples, a procedural blank was included, which was treated as the same as the soil sample to check the procedural performance and matrix effects. The surrogate standard recoveries in samples ranged from 85~121%.

Multivariate analysis
Some descriptive statistical parameters, including mean, median, minimum and maximum concentrations, skewness, variation coefficient, etc., were carried out by SPSS 19.0 and Excel 2016. The spatial distribution of heavy metals and PAHs was mapped using the inverse distance weighting method by Arc-GIS 10.2 interpolation analysis. Variation coefficient (VC) was used to reflect the degree of discrete distribution of different contaminants, as well as the impact of human activities on the environmental distribution of pollutants. Skewness was utilized to reflect different distributions of the contaminants (Yongming et al., 2006). To assess the contamination of heavy metals in the soil, the pollution index (PI) was adopted.
Multivariate analysis methods were used to assess the interaction and co-contamination of two groups of pollutants in the environment. Bivariate analysis (Pearson correlation) was used to determine the correlations between heavy metals and PAH distributions. Principal component analysis (PCA) was also used to identify the differences in the concentrations and distributions between different groups of pollutants. A redundancy analysis (RDA) with Monte Carlo was performed to determine the multivariate relationship between different land-use types and the co-contamination of PAHs and heavy metals in the region. Data processing and graphing for PCA and RDA were done using Origin 2019 software.

Heavy metal contamination
The levels of Cr, Ni, Cu, Zn, As, and Pb in the study area were shown in Table 1 -1995), the heavy metal content of some sites in this study exceeded the standard. The sampling points with Cr and Ni concentrations above the standard accounted for 13.6% and 27.2% of the total sampling points, respectively, with the exceeding points concentrated primarily in Fushun. In addition, concentrations of Cu and Zn exceeded the limited in 31.8% and 12.1% of sampling sites, which were mainly found in Shenyang. The concentrations of Pb and As in the sample sites were all below the permissible limit.
Skewness values indicated that the distribution of Ni concentrations was normal in the sampling sites, while the distribution of the other metal elements (Cr, Cu, and Pb) was positively skewed towards the higher concentrations. The skewness values showed that some sites have been highly contaminated by heavy metals. This can also be supported by the fact that the mean concentrations of these metals are much higher than their median concentrations. Among all metal elements, VCs of Ni and As were < 40%, while VCs of Zn, Cr, Cu, and Pb were > 67%. It was reported that elements dominated by a natural source had low VCs, while those elements affected by anthropogenic sources had relatively higher VCs (Yongming et al., 2006). Combined with skewness results, we could speculate that Ni and As were mainly from natural sources, and it could be further confirmed by the similar concentrations of Ni and As were found in the study and the background concentration. We could also speculate that the distribution of Cr, Cu, Zn, and Pb was affected by humans. Another study in Xi'an also showed that these four metal elements had higher levels in urban areas, which were from industrial and traffic sources (Yongming et al., 2006).

PAH contamination
The concentration of PAHs was shown in Table S3. The concentrations of ∑ 23 PAHs ranged from 29.51 to 29,601.09 μg/kg, with a mean concentration of 3906.68 μg/kg. As a background reference, the arithmetic means of ∑ 21 PAHs in this region was 726.94 μg/kg, which indicated that some of the sampling sites were highly contaminated. The mean concentration of PAHs of different numbers of rings in surface soil is 158.08 μg/kg (2 rings), 1555.48 μg/ kg (3 rings), 1649.90 μg/kg (4 rings), 419.91 μg/kg (5 rings), and 123.30 μg/kg (6 rings). These results suggested that the 3 ring and 4 ring PAHs were the predominant of ∑ 21 PAHs in the region. Regarding all the PAH individuals, there were six PAHs (including BaA, Chry, BbF, BaP dBaAnt, and InP), which were considered carcinogenic PAHs by USEPA, exceeded the permissible concentration (Cao et al., 2017;Korre., 1999;Wang et al., 2018). The results revealed that there was a higher health risk in some particular sites in the studied area. Significant differences were found in the skewness and VCs between PAHs and heavy metals. All the skewness values were > 2, with upper values of > 5, which indicated a significantly abnormal distribution. Similarly, the VC values were above 150%, indicating that there was a strong influence of human activities on the distribution of PAHs. The results above also demonstrated that the distribution patterns of heavy metals and PAHs were different, which could be attributed to anthropogenic sources. The pollutions of heavy metals were a combination of natural and anthropogenic results, while PAHs pollution was mainly from anthropogenic sources (Cao et al., 2017;Korre., 1999;Wang et al., 2018).
To figure out the pollution of heavy metals and PAHs in other areas, we collected data from other studies (shown in Table S4). The main sources of pollution in this area were industrial activities, sewage irrigation, etc. In comparison, our results were higher than the concentrations of PAHs, Cr, Ni, and Cu in the soil of Shenyang in 2006 (Song et al., 2006), demonstrating the accumulation of heavy metals and PAHs in recent years. The concentrations of PAHs in an industrial city in southern Italy (Sprovieri et al., 2007) ranged from 9 to 31,774 μg/kg, which was similar to our results (29.51~29,601.09 μg/kg). The concentration of PAHs in our study was similar to that in urban areas such as Central Serbia (38~3136 μg/kg) (Stajic et al., 2016) and Beijing (181~4092 μg/kg) , while the concentrations of heavys metal were higher in our study. However, higher concentrations of PAHs were found in soils collected from a coking plant in Chenzhou (4520~8210 μg/ kg) (Zhu et al., 2012), and a city in northeast India (13,480~86,300 μg/kg) with oil and gas drilling near the city. Overall, the pollutions of PAHs and heavy metals in this study were similar to that in industrial cities (such as industrial cities in Italy and Chenzhou), but higher than that in some urban areas (such as Beijing and southern Serbia). Therefore, our results suggested that there might be complex interactions and associations between heavy metals and PAHs in this area (Shen et al., 2005).

The distribution of heavy metals and PAHs
The pollution index for heavy metals detected in surface soil were shown in Fig. 2. The PI was defined as the ratio of the heavy metal concentration in the study to the geometric means of background concentration (BC) of the corresponding metal of the studied area. The PI of each metal was calculated and classified as either low (PI ≤ 1), medium (1 < PI ≤ 3), or high (PI > 3) (Chen et al., 2005). Overall, Ni and As had lower contamination levels, and most of the PIs were ≤ 1, which was similar to the analysis above. For Cr, Cu, Zn, and Pb, the PI for most of the samples were either low (PI ≤ 1) or medium (1 < PI ≤ 3). A few samples were found to have high PI values, such as Cr, Ni, and Cu for NO. SF08 and NO. SF60 at Shenfu New District, and Cu, Z, and Pb for NO. SS39 in Shenyang, whose PI values were > 10, indicates high contaminations at these sites. Compared with other regions, Fushun had relatively higher PI values of Cr and Ni, and Shenyang had higher PI values of Zn and As. This could be explained by the different industrial types in the two regions. Fushun is a resource-based city with coal, chemical, and metallurgical industries, which have been identified as major sources of Cr and Ni in many studies (Chen et al., 2005;Wu et al., 2019;Yongming et al., 2006). In addition, Han also found that the coal mining in Fushun affected the distribution of Ni (Han et al., 2012). Shenyang has comprehensive industries and millions of vehicles, and the industrial sources and traffic sources may affect the distribution of Cr, Cu, Zn, and Pb in this region. In some samples collected from Fushun and Shengyang, relatively higher PI values were found for Ni and As, respectively, indicating that anthropogenic sources were the major sources in some sampling sites.
The concentration profile of PAHs in the 66 soil samples collected from different areas was shown in Fig. 3. Background concentrations of PAHs were displayed as a blue dotted line in the figure. Concentrations of PAHs in most of the samples were higher than the background concentrations, and the concentrations in Shenyang and Fushun were higher than that in Shenfu New District. The results demonstrated the influences of urbanization time on the contamination of PAHs in different areas. Shengyang and Fushun start the urbanization process decades ago, while Shenfu New District is a newly developed city. Meanwhile, the average concentrations in Shenyang were higher than that in Fushun. The accumulations of 3 ring and 4 ring PAHs in the city were observed significantly, with the concentrations of some sites exceeding 15,000 μg/kg. Relatively high concentrations of PAHs were found in several sites, such as NO. SS29, NO. SS32, NO. SS39,and NO. SS42 in Shenyang,and NO. FS10,NO. FS18,and NO. FS21 in Fushun. Meanwhile, high levels of heavy metals were also found in these samples, such as NO. SS32, NO. SS39, NO. SS42, and NO. FS18, which indicated the combined pollution of the two groups of pollutants in these areas.
The spatial distribution patterns of heavy metals and PAHs in the region were plotted by ArcGIS 10.2 (ESRI) and were shown in Fig. 4 and Fig. 5. Samples that exceeded the permissible concentrations were marked by red circles in Fig. 4. High concentrations of Cr were mainly distributed in the north and south of Fushun, but no accumulation was found in the urban area of Shenyang and Fushun. Similarly, higher concentrations of Ni were also found in urban areas of Fushun. For the concentrations of Cu, above permissible concentrations were found in a large number of urban sites in Shenyang and Fushun, indicating the high pollution in the urban area. The other heavy metals (Pb, As, and Pb) were mainly distributed in the urban areas of Shenyang, and concentrations of As and Pb were all below the permissible concentration. Figure 5 showed the concentration distribution of ∑ 21 PAHs, where high concentrations were found in the urban areas of Shenyang and Fushun City. Overall, there was substantial evidence for the impact of industrial and population aggregation on the accumulation of heavy metals and PAHs. In addition, high concentrations of Cu, Zn, As, and Pb were also found in sites with high PAH concentrations, and there was a high possibility of combined pollution in these areas. In this regard, we used multivariate analysis to analyze the correlations between concentrations of heavy metals and PAH in different sampling sites.

Correlation analysis
Pearson correlation analysis was utilized to assess the relationships between heavy metals and PAHs ( Table 2). The Pearson correlation coefficients (r) were-0. 003, 0.168, 0.498, 0.677, 0.556, and 0.733 for Cr, Ni, Cu, Zn, As, Pb, and PAHs, respectively, in which the correlations for Cu, Zn, As, Pb, and PAHs showed the significance at 0.01 level (P < 0.01). There were three possible explanations for the significant correlation (Wang et al., 2018): (1) the PAHs and heavy metals in the sampling area might have originated from the same sources; (2) the distribution patterns of PAHs and heavy metals are similar to each other; and (3) adsorption competition and synergistic reaction between PAHs and heavy metals might co-exist in the soil. It can be therefore speculated that the sources of Cu, Zn, As, Pb, and PAHs were similar, and there were synergistic pollutions in the soil. Note that the correlation coefficient between Cr and Ni was 0.755, indicating that the two substances have a very similar distribution, sources, and pollution behavior in the soil. Correlation analysis also showed that Cu, Zn, As, and Pb are significantly correlated with PAH, suggesting the necessity of assessing the combined pollution of the two groups of pollutants.
The correlation coefficients between heavy metals and PAHs with different numbers of rings were shown  Table 2. The correlation coefficients ranged from 0.304 to 0.775, with significant correlations among these pollutants (P < 0.01). Stronger correlations were found between metals (Cu and Pb) and heavier PAH, indicating that the associations between Cu/Pb and high-ring PAHs were closer, and co-contamination among them was more significant. This weight-based trend was probably because of the physicochemical properties of different PAHs that the high molecular weight PAHs were more stable in the soil, and the synergistic effects and combined pollution with heavy metals were more significant. The correlation between Zn/As and PAHs is not associated with the number of rings, and the correlation coefficients were between 0.415 and 0.651 across PAHs with a different number of rings. Combined with the spatial distribution, we could speculate that combined pollutions of Cu, Zn, As, Pb, and PAHs existed in urban areas of Shenyang and Fushun.

Principal component analysis
Principal component analysis (PCA) is a common multivariate statistical method used in scientific research and has been widely used to reduce variable components and to extract a small number of latent factors for analyzing relationships among the observed variables (Praveena et al., 2012). It can reveal the internal relations of soil pollutant data and describe the main process of soil pollutant law (Korre., 1999), and also be used to conduct a quantitative assessment of soil pollution in areas where soil pollutants are complex or have insufficient monitoring data (Yongming et al., 2006). Therefore, the PCA was utilized to assess the combined pollution and to investigate the inherent relationship among pollutants.
SPSS for Windows(19.0, SPSS Inc, USA) was used for principal component analysis. Data was normalized before PCA, to overcome the huge variation in the concentrations of pollutants in different sites. Meanwhile, all principal factors extracted from the variables were retained with eigenvalues > 1.0 (Yongming et al., 2006). The principal component factor loads were shown in Fig. 6 and Table S5, which showed clear relationships between heavy metals and PAHs.
Three components were obtained from the PCA, accounting for 88.62% of the total variance, which could explain most of the data and have high credibility. The PC1 was dominated by Cu, Zn, As, Pb, Acy, Ace, Fl, Debt, Phe, Ant, Fla, Pyr, BaA, Chry, BbF, BkF, BaP, Pery, InP, dBaAnt, and BghiP (mainly 3~6 ring PAHs), accounting for 65.48% of the total variance, reflecting the homology of these pollutants in the samples as well as the complicated combination situation; the PC2 was dominated by NaP, 2-Met, 1-Met, and Ret (2 ring PAHs except Ret), accounting for 14.95% of the total variance, containing information of 2-ring PAHs; the PC3 accounts for 8.19% of the total variance, which contained information of some heavy metal elements, including Cr, Ni, and Cu. At the same time, it was worth noting that Retene had a different behavior in the loading results, and further analysis was needed to identify the associated behaviors.
In the loading plot, the distance between the variable and the origin indicated the loading value of a given factor. The variable located far from the origin had a larger loading value and the variable located near the origin had a smaller loading value. In Fig. 6 and Table S5, we found that Cu, Zn, As, Pb, and PAHs (3 or higher ring PAHs) were the main loads, which had significant correlations with each other in the statistical analysis, indicating the combined pollution in this region. In addition, the 2 ring PAHs (NaP, 2-Met, 1-Met), which had the second-highest contribution to the loading values were not correlated to the distribution of heavy metals. Thus, it can be inferred that the interactions between 2 rings of PAHs and heavy metals were not obvious in this region. Cr and Ni were located near the origin, and they were not significantly correlated with other substances. This was consistent with the results of the previous correlation analysis. Because of the properties and interactions of PAHs and heavy metals, low-ring PAHs could be easily decomposed by microorganisms in the soil, which resulted in their short half-lives and their relatively low concentration in the soil (Table 2) (Wei-gen., 2010). Therefore, the relationships between heavy metals and PAHs depended on the number of rings (i.e., 2 ring PAHs showed a different behavior from other PAHs). On the other hand, an important way for the interaction between heavy metals and PAHs in the soil was through the cation-π bond (Dougherty, 2013). PAHs with a higher π-electron density (i.e., pyrene > phenanthrene > naphthalene) had more opportunities to establish associations with heavy metals (Sushkova et al., 2019). Combined with the distribution patterns and contamination levels of the two groups of pollutants in this region, it could be inferred that the co-contamination in the region mainly happened among Cu, Zn, As, Pb, and 3~6 ring PAHs.

RDA analysis
Some physical and/or chemical properties, such as pH value and organic matter would have a significant impact on the behavior and interaction of pollutants (Sushkova et al., 2019). Some studies showed that the PAHs were mainly in soil organic matter (Nam et al., 2008). Therefore, we introduced soil organic matter data as an environmental variable to investigate its role in the interaction between heavy metals and PAHs. Redundancy analysis (RDA) is a commonly used ranking analysis tool for environmental ecology statistics, which can be used to reveal the intrinsic linkages and interactions between species and environmental factors, and has also been applied to analyze the environmental behavior of pollutants in recent years (Zhang et al., 2017). Based on this, RDA analysis was applied (Canoco for Windows 4.5) The results of detrended correspondence analysis (DCA) showed that the lengths of the first ordination gradient were less than 3, RDA was adopted to examine the correlations between the PAHs and environmental variables (e.g., SOM, Cr, Ni, Cu, Zn, As, Pb). The results of RDA for the three cities were shown in Fig. 7. In the left panel (Fig. 7), the first and second axes of the RDA for Shenyang accounted for 80.3% and 3.4%, of the total variance respectively. Soil organic matter contains various functional groups that can adsorb heavy metals and PAHs through hydrogen bonds, van der Waals forces, and coordination bonds (Cao et al., 2017). As a result, both heavy metals and PAHs can be found in solid organic matter, and complex interactions may occur under the influence of microorganisms . In the RDA diagram, the arrows between the SOM, heavy metals, and different-ring PAHs pointed in a similar direction, at an acute angle < 90°. In Shenyang, each substance was positively correlated, the angle of Zn/As with 2~6 rings, and Cr/Cu with 4~6 ring PAHs was < 30°, showing strong correlations among them. The length of the arrow represented the proportion of the variance explained, and Zn, As, Pb, Cr, Cu, Ni, and SOM explained 66.0%, 47.8%, 44.8%, 42.4%, 40.2%, 29.4%, and 44.6% of the variance variables respectively. Since heavy metals and SOM were commonly used as environmental variables and there was a positive correlation between the pollutants, this ratio could indicate the probability and degree of combined pollution of different heavy metal elements and PAHs. Overall, the cross-combination of the two groups of pollutants in the figure also indicated that complex combined pollution happened in soils collected from Shenyang.
Compared to Shenyang, different results were found in the other cities (right panel in Fig. 6). The first and second axes of the RDA plot explain 25.5% and 1.9% (Shenfu New District), 38.5%, and 1.6% (Fushun) of the total variance, respectively. Different from the positive correlations found in Shengyang, Cr, Ni, and Cu were negatively correlated to some other elements in Shenfu New District, and SOM, Pb, Zn, and As were positively correlated to PAHs, suggesting that combined pollution mainly happened among Pb, Zn, As, and PAHs. In Fushun, the angles among As and different rings of PAHs are close to 90°, suggesting there was no correlation between As and PAHs. In addition, there were positive correlations between Pb/Ni and 2 ring PAHs. The results of RDA showed that there were differences among the three regions in the distribution patterns and interactions of heavy metals and PAHs, and the different results also indicated the impact of two industrial cities on this region.

Conclusion
The skewness values of the heavy metals in the soil reveal that Cr, Cu, Zn, and Pb accumulation is primarily caused by human activity, whereas Ni and As accumulation is primarily caused by natural sources. The distribution of heavy metals is significantly influenced by the city's industrial character. In the study area, 3-ring and 4-ring PAHs make up the majority of the PAHs. The distribution of PAHs exhibits a stronger human impact than heavy metals, as evidenced by the difference between PAH and heavy metal skewness results. Cu, Zn, As, and Pb have a stronger relationship with PAH than other heavy metals, especially PAH with more than three rings. Further verification with PCA reveals that the synergistic relationship between heavy metals and PAHs is dependent on the number of rings of PAH, that is, co-pollution primarily occurs between Cu, Zn, As, Pb, and PAH with 3~6 rings, implying that these four types of heavy metals must be given special consideration when evaluating synergistic pollution. The RDA results show that the distribution patterns and interactions of heavy metals and polycyclic aromatic hydrocarbons differ between the three research cities. The complexity of compound pollution in urban soil in Shenyang is significantly higher than in Fushun and Shenfu New District, demonstrating that the urbanization process and industrial type are important factors influencing the type of compound pollution.