Vertical Distribution Characteristics and Source Analysis of Heavy Metals in Near-Surface Urban Dust in Central Beijing
ChenSu1
WenjiZhao1✉Email4973@cnu.edu.cn
JieDong1
YiXueZhong1
ZhiQiangYang1
1
A
A
College of Resource Environment and TourismCapital Normal UniversityChen Su 1,, Wenji Zhao1,*, Jie Dong 1,YiXue Zhong 1, ZhiQiang Yang 1
1College of Resource Environment and Tourism, Capital Normal University
* Correspondence:
Wenji Zhao
4973@cnu.edu.cn
Keywords:
Vertical distribution
APCS-MLR model
Heavy metals
roof dust; Urban dust
Abstract
Urban dust serves both as a sink for heavy metals and a secondary source; however, its vertical distribution and source composition remain insufficiently explored. This study investigated the heavy metal concentrations in rooftop and ground dust samples collected from Haidian District, Beijing. Using Inductively Coupled Plasma Mass Spectrometry (ICP-MS) and receptor models (PCA and APCS-MLR), we analyzed the spatial variation, pollution levels, ecological risks, and source apportionment of eight heavy metals (V, Cr, Ni, Cu, Zn, As, Cd, Pb). The results indicated that rooftop dust exhibited generally higher concentrations and ecological risks, with cadmium identified as the most critical pollutant. APCS-MLR analysis revealed that rooftop dust was predominantly influenced by an unidentified source (67.4%), whereas ground dust was mainly affected by combustion sources (42.4%) and traffic emissions (18.2%). This vertical heterogeneity suggests that rooftop dust primarily reflects regional atmospheric deposition, while ground dust captures local pollution signals. These findings underscore the necessity for three-dimensional environmental monitoring and targeted pollution control strategies in urban areas.
A
Introduction
With the continuous advancement of urbanization and the acceleration of industrialization, urban spaces are increasingly expanding vertically. High-rise buildings are becoming more prevalent, and daily human activities, both in work and life, are progressively extending upward into vertical urban spaces (Li et al 2015). Consequently, the vertical distribution characteristics of atmospheric pollutants have attracted growing attention. This distribution not only affects the exposure levels of people at different heights but also has important implications for ground-level air quality (Jo et al., 2006; Ma et al., 2007) and public health.
Urban deposited dust comprises a mixture of various pollutants and is closely linked to air and water environmental pollution (Zhao et al., 2023). Under the influence of natural forces such as wind, water, and gravity, these pollutants adhere to and accumulate on surfaces at different urban elevations (Zou et al., 2017; Liu et al., 2012). As both a sink and a source of pollutants (Yadav et al., 2018), urban dust typically contains heavy metals at concentrations higher than local soil background levels (Lu et al., 2014). Given their persistence, non-degradability, and potential for bioaccumulation, heavy metals pose significant threats to human health and ecosystems (Roy et al., 2022).
Previous studies have indicated that heavy metals in road dust may originate from soil resuspension (Zhang et al., 2010), traffic emissions (Shi et al., 2018; Mummullage et al., 2016; Ali-Taleshi et al., 2022), degradation of construction materials (Liu et al., 2007), and industrial activities. Identifying pollution sources is a critical prerequisite for effective environmental pollution prevention and control (Duan et al., 2023). Currently, both qualitative methods—such as Principal Component Analysis (PCA) and Factor Analysis (FA)—and quantitative receptor models—such as Positive Matrix Factorization (PMF) (Ramírez et al.,2020), Absolute Principal Component Scores–Multiple Linear Regression (APCS-MLR) (Wang et al., 2021), and UNMIX (Song et al., 2006)—are widely applied for source apportionment (Duan et al., 2024).
However, existing research has primarily focused on easily accessible ground-level areas such as roads (Zhao et al., 2021;Roy et al., 2022), industrial zones (Lu et al., 2020), and school surroundings (Fan et al., 2021), while relatively little attention has been paid to heavy metal pollution in high-rise environments, such as rooftop dust. Li Dunzhu et al. (Li et al., 2015) reported that the average heavy metal concentrations in rooftop dust are significantly higher than those in road dust. Compared to surface dust, rooftop dust is less affected by rainfall, pedestrian activity, and traffic emissions, making it a more reliable indicator of atmospheric dry deposition and a more accurate reflection of airborne particulate pollution levels (Wang et al., 2021).
Therefore, investigating the characteristics and sources of heavy metals in rooftop dust is essential for understanding the vertical transport and accumulation mechanisms of pollutants, and for addressing the current gap in three-dimensional urban pollution studies.
This study selects Haidian District, a representative urban area in Beijing, as the research site. Dust samples were collected from both ground level and high-rise rooftops. Heavy metal concentrations were measured, and the APCS-MLR model was employed for source apportionment. The study analyzes pollution levels, source composition, and ecological risks associated with dust at different elevations, and explores the vertical distribution characteristics of heavy metal pollution in urban environments. The findings are expected to contribute to the improvement of three-dimensional atmospheric pollution monitoring systems and provide a scientific basis for refined urban environmental management.
Materials and Methods
Overview of the Study Area
Haidian District is located in the western and northwestern part of urban Beijing, ranging from 39°53′ to 40°09′N and 116°03′ to 116°23′E. The area experiences a typical temperate humid monsoon climate, characterized by hot and rainy summers dominated by southeasterly winds, and cold, dry winters with prevailing northwesterly winds. Haidian serves as a national hub for high-tech industries, higher education, and scientific research institutions, with a permanent population of approximately 3.13 million. The study site is situated within the main campus of Capital Normal University, located along the West Third Ring Road (Fig. 1), representing a central urban setting. The sampling locations, positioned within the urban core, can to some extent reflect the general environmental conditions of central Beijing.
Fig. 1
Geographical distribution of sampling points
Sample Collection and Preparation
From April to September 2021, a total of 84 dust samples were collected from 14 sites within the study area, comprising 42 road dust and 42 rooftop dust samples (Fig. 1). At each site, a 1 × 1 m wooden frame was deployed, with the base lined by an antistatic, non-adhesive polyolefin (PO) film. This film, primarily composed of polyethylene, has a smooth surface that minimizes particle adherence, enabling the natural deposition of dust particles without external contamination. To ensure structural stability, the frames were secured using concrete bricks, protecting them from wind or inclement weather. Following deposition, accumulated dust on the PO film was carefully collected using a clean plastic brush and transferred into sterile envelopes.
Road dust samples were obtained by sweeping the surface with a soft plastic brush and collecting particles with a plastic scoop. Samples were placed in sealed polyethylene bags and transported to the laboratory. All tools were thoroughly cleaned between sampling events to prevent cross-contamination.
In the laboratory, samples were air-dried in aluminum containers. Non-dust materials, including plant debris, gravel, and insect remnants, were manually removed using plastic tweezers. The remaining dust was homogenized with a mortar and passed through a 100-mesh nylon sieve. For chemical analysis, approximately 0.2 g of each sample was digested using a mixed acid solution (HNO₃–HCLO₄–HF). Concentrations of eight heavy metals—vanadium (V), chromium (Cr), nickel (Ni), copper (Cu), zinc (Zn), arsenic (As), cadmium (Cd), and lead (Pb)—were quantified using inductively coupled plasma mass spectrometry (ICP-MS), following standard protocols (Ministry, 2015; Wang et al., 2015).
To ensure analytical quality, each batch of samples included three procedural blanks and certified reference materials (GSS-1, GSD-12). Quality control and assurance procedures were implemented throughout the digestion and analysis process to validate the accuracy and precision of the measurements.
Evaluation Methods
Assessment of Heavy Metal Pollution in Dust
To evaluate the pollution levels of heavy metals in rooftop and ground dust, the potential ecological risk index (RI) proposed by Hakanson (Hakanson ,1980) was adopted. The calculation is as follows:
A
Where RI (ecological risk index) is the sum of ecological risk factors (); n is the heavy metal species studied in this paper, which is 8;
is the monomial potential ecological risk factor for i species;
is the toxic response factor for a given metal;
is the contamination factor of i and numerically equal to PI value; Ci refers to the concentration of the metal in the dust samples, while Bi represents the background concentration of the metal in the environment. According to the calculation of
and RI, the ecological risk levels caused by metal elements in dust are described:
≤40 or RI≤150 indicates low potential ecological risk; 40<
≤80 or 150<RI≤300 indicates moderate potential ecological risk; 80<
≤160 or 300<RI≤600 indicates considerable risk; 160<
≤320 or 600<RI indicates high risk;
>320 indicates extreme risk.
The Pollution Load Index (PLI) was proposed by Tomlinson et al. (Tomlinson et al., 1980) as a method to evaluate the overall level of heavy metal pollution at a given site by integrating the contamination levels of multiple elements. It provides a more comprehensive reflection of heavy metal contamination. The calculation formulas are as follows:
Where Ci is the measured concentration of a heavy metal in the dust sample, Cₙ is the corresponding background concentration in the environment, and PLI represents the overall pollution load index for a specific sampling site. The pollution level is classified based on the value of PLI as follows: no contamination (PLI ≤ 1), low level of contamination (1 < PLI ≤ 2), middle level of contamination (2 < PLI ≤ 3), and very high level of contamination (PLI > 3)(Li et al., 2015).
Correlation Analysis (CA)
Correlation Analysis (CA) is a multivariate statistical method used to measure the strength of the relationship between two or more interrelated variables. By analyzing the correlations among different heavy metal elements, it is possible to determine whether these metals may share the same or similar sources (Shi et al., 2022).
APCS-MLR
The APCS–MLR model was first proposed by Thurston and Spengler (Thurston et al., 1985) as a method for source identification and quantification of contributions from various pollution sources, based on Principal Component Analysis (PCA) and Multiple Linear Regression (MLR). In this approach, conventional PCA is used to generate Absolute Principal Component Scores (APCS), which are then used as independent variables, while the concentrations of heavy metals are taken as dependent variables to perform MLR and determine the contribution rates of different pollution sources. The calculation procedure is as follows:
Standardization and Principal Component Score Extraction:
Calculation of Factor Scores for Zero-Concentration (Artificial) Samples:
Where
is the standardized concentration value of metal k at a sampling point,
is the measured concentration,
is the arithmetic mean, δ is the standard deviation,
is the loading of the jth principal component,
is the principal component score,
is the standardized value when the concentration of metal iii is zero,
is the score coefficient of the jth principal component for metal i,
is the factor score of component j for the zero-concentration sample.
Calculation of APCS and Source Contribution Using MLR:
Where
is the absolute principal component score of the jth component at sample k,
is the measured concentration of metal i,
is the regression coefficient representing the contribution of source m to metal i,
is the constant term (representing unidentified sources), n is the number of components.
Source Contribution Rate Calculation:
The contribution rate of source mmm to metal iii is calculated as:
The contribution rate of unidentified sources is given by:
Where
is the contribution rate of source mmm to metal i,
is the average APCS for all samples of metal i.
Results and Discussion
Heavy Metal Concentrations in Ground and Rooftop Dust
As shown in Table 1, the concentrations of heavy metals in rooftop dust are generally higher than those in ground dust. This trend is particularly evident for elements such as Cu, Zn, Cd, and Pb, which are closely associated with traffic emissions and industrial pollution, indicating significant enrichment. For instance, the average concentrations of Zn and Cd in rooftop dust are 336.06 mg/kg and 0.70 mg/kg, respectively, which are considerably higher than those in ground dust (238.22 mg/kg and 0.52 mg/kg). These values are 3.28 and 9.43 times higher than the corresponding background levels in Beijing soils (Chen et al., 2004), suggesting a clear vertical enrichment trend in the urban environment.
In contrast, elements such as V, Cr, and Ni show relatively similar concentrations in both rooftop and ground dust samples, with some values even falling below the background levels. This indicates that these elements may primarily originate from natural sources or have limited anthropogenic input.
The coefficient of variation (CV) is a normalized measure reflecting both the regional heterogeneity of soil heavy metals (Li et al.,2019) and the degree of dispersion in their probability distribution. A higher CV indicates greater dispersion, often reflecting stronger anthropogenic influence (Lu et al.,2023). According to the CV analysis (Wilding et al., 1985), ground dust shows higher overall variability in heavy metal concentrations compared to rooftop dust, especially for Cu, Zn, Cd, and Pb, which exhibit high variability. This suggests that their spatial distribution is strongly affected by local human activities, such as traffic density, construction, or proximity to pollution sources.
In contrast, most heavy metals in rooftop dust show low to moderate variability, implying less disturbance and more spatial uniformity, which better reflects the characteristics of regional atmospheric dry deposition.
In summary, heavy metal concentrations in ground dust are more strongly influenced by diverse local sources, whereas rooftop dust likely reflects regional atmospheric emissions and dry deposition processes, indicating a background pollution signal. In the following sections, receptor models will be employed to further identify specific pollution sources for both types of dust.
Dust | Element | V | Cr | Ni | Cu | Zn | As | Cd | Pb |
|---|---|---|---|---|---|---|---|---|---|
Ground | Min/mg.kg− 1 | 51.58 | 36.65 | 12.22 | 5.55 | 40.26 | 4.68 | 0.11 | 18.16 |
Max/mg.kg− 1 | 78.60 | 85.80 | 37.65 | 90.20 | 462.86 | 13.63 | 0.86 | 69.47 | |
Mean/mg.kg− 1 | 64.22 | 65.30 | 27.37 | 40.38 | 238.22 | 8.25 | 0.52 | 45.57 | |
SD/mg.kg− 1 | 9.91 | 13.60 | 6.86 | 25.57 | 128.64 | 2.84 | 0.27 | 17.28 | |
CV% | 0.15 | 0.21 | 0.25 | 0.63 | 0.54 | 0.34 | 0.52 | 0.38 | |
Skewness | 0.11 | -0.85 | -0.78 | 0.46 | 0.09 | 0.82 | -0.37 | -0.28 | |
Kurtosis | -1.46 | 0.04 | 0.26 | -0.88 | -1.06 | -0.50 | -1.43 | -1.27 | |
Rooftop | Min/mg.kg− 1 | 61.16 | 60.07 | 25.24 | 46.92 | 252.02 | 8.86 | 0.55 | 45.91 |
Max/mg.kg− 1 | 84.48 | 84 | 37.45 | 93.48 | 529.64 | 13.92 | 0.86 | 100.22 | |
Mean/mg.kg− 1 | 72.71 | 74.79 | 31.53 | 65.44 | 336.06 | 10.55 | 0.70 | 59.81 | |
SD/mg.kg− 1 | 5.25 | 6.98 | 3.17 | 13.75 | 66.93 | 1.34 | 0.08 | 13.50 | |
CV% | 0.07 | 0.09 | 0.10 | 0.21 | 0.20 | 0.13 | 0.12 | 0.23 | |
Skewness | 0.00 | -0.84 | -0.06 | 0.61 | 1.89 | 1.41 | -0.25 | 2.23 | |
Kurtosis | 2.39 | 0.20 | 0.20 | -0.22 | 5.21 | 2.17 | 0.52 | 6.28 | |
Beijing soils/mg.kg− 1[30] | 79.2 | 68.1 | 29 | 23.6 | 102.6 | 9.7 | 0.074 | 25.4 | |
Characteristics of Heavy Metal Pollution in Ground and Rooftop Dust
Potential Ecological Risk Assessment
The statistical results of potential ecological risk at different heights are shown in Fig. 2. As depicted, the potential ecological risk index (Erⁱ) values for elements such as V, Cr, Ni, Cu, Zn, As, and Pb in both ground and rooftop dust are all below 40, indicating a low ecological risk.
However, the Erⁱ values of Cd in both ground and rooftop dust are significantly elevated, with mean values of 210.92 and 283.03, respectively. For ground dust, 21.43% and 7.14% of the sampling sites fall into the moderate and considerable risk categories, while 57.14% and 14.29% fall into the high and very high-risk categories. In rooftop dust, 85.71% of the sites are classified as considerable risk, and 14.29% as high risk. These results indicate that Cd is the dominant contributor to ecological risk in both ground and rooftop dust.
According to Table 2, which presents the comprehensive potential ecological risk index (RI) at different heights, the RI values of rooftop dust are generally higher than those of ground dust. Among the ground dust samples, 28.57% fall into the low-risk category, 28.57% into moderate risk, and 42.86% into the considerable risk category. In contrast, rooftop dust samples show 14.29% in moderate risk, and 85.71% in considerable risk.
Moreover, Cd contributes far more to the total RI than any other heavy metal, underscoring its importance and indicating that Cd should be a key focus in future studies on urban heavy metal pollution.
Dust | RI Range | Mean | Proportion of Potential Ecological Risk % | ||||
|---|---|---|---|---|---|---|---|
low | moderate | considerable | high | extreme | |||
Ground | 79.28-397.22 | 247.94 | 28.57 | 28.57 | 42.86 | 0 | 0 |
Rooftop | 263.74–402.10 | 332.29 | 0 | 14.29 | 85.71 | 0 | 0 |
Fig. 2
Assessment of potential ecological risks of dust accumulation at different heights
Assessment of Heavy Metal Pollution Levels
As shown in Fig. 3, based on the background values of Beijing soils, the Pollution Load Index (PLI) of heavy metals at different heights range from 0.54 to 2.20 for ground dust and from 1.69 to 2.30 for rooftop dust. Among the ground dust samples, 21.43% of sites are classified as unpolluted, 64.29% as slightly polluted, and 14.29% as moderately polluted. In contrast, 50% of the rooftop dust samples fall into the slightly polluted category, and the remaining 50% into the moderately polluted category.
Overall, except for sampling sites 1, 2, and 8—where the ground PLI values slightly exceed those of rooftop dust, the average PLI values for rooftop dust are higher than those of ground dust across most sampling sites.
Based on the mean Contamination Factor (CF) values of each metal, the pollution order for ground and rooftop dust is as follows:
Ground dust: Cd > Zn > Pb > Cu > Cr > Ni > As > V
Rooftop dust: Cd > Zn > Cu > Pb > Cr > As > Ni > V
For ground dust, the CF values of Cr, Ni, As, and V are all less than 1, indicating concentrations below background levels and thus no pollution. Similarly, only V in rooftop dust shows a CF value less than 1. In both ground and rooftop dust, Cd exhibits the highest CF values, indicating that Cd is the dominant pollutant in the study area.
This elevated Cd pollution may be attributed to the location of the sampling area near Beijing’s Third Ring Road, which is heavily influenced by traffic activities (Cheng et al., 2013.).
Fig. 3
Heavy metal pollution load index of different heights of dust accumulation
Ground and Rooftop Heavy Metal Source Analysis
Correlation Analysis
Pearson correlation coefficients were used to analyze the relationships between heavy metals in ground and rooftop dust, and the results are visualized using heatmaps. The strength of the correlation coefficients can be used to infer whether certain metals share similar geochemical behaviors or common sources. In general, the stronger the correlation, the higher the likelihood of a shared source (Li et al., 2020).
As shown in Fig. 4, V, Ni, and As exhibit strong mutual correlations in ground dust, indicating possible co-occurrence or similar sources. In rooftop dust, however, V only shows strong correlations with Cr and As, and a negative correlation with Pb. Additionally, Ni and As still maintain a strong correlation in rooftop dust. This variation suggests that the correlation between V, Ni, and As observed at ground level may diminish in rooftop samples due to changes in influencing factors with elevation.
Other examples of such variations include:
Cr and Ni are strongly correlated in ground dust, while Cr and Zn show a stronger correlation in rooftop dust.
Cu exhibits strong correlations with Ni, As, Cd, and Pb in ground dust, whereas in rooftop dust, Cu shows only a negative correlation with Cd.
Zn is strongly correlated with Cd and Pb in ground dust, while rooftop dust it correlates more with Cr and Pb.
These differences likely result from the more complex and diverse influences on ground dust, including localized anthropogenic activities, which alter the inter-element relationships observed at different heights.
Fig. 4
Correlation coefficient between heavy metals with different heights of dust
PCA
Standardized data processed using SPSS 26 were subjected to the Kaiser-Meyer-Olkin (KMO) test and Bartlett’s test of sphericity. The KMO values for ground and rooftop dust were 0.686 and 0.592, respectively (both above the threshold of 0.500), and the p-values for Bartlett’s test were both less than 0.05. These results indicate that the data are suitable for factor analysis.
Principal Component Analysis (PCA) extracted three principal components (PCs) for both ground and rooftop dust, with cumulative variance contributions of 94.09% for ground dust and 82.74% for rooftop dust, indicating that most of the information in the original dataset is retained.
In ground dust, PC1 shows high loadings for Zn (0.953), Cd (0.908), and Pb (0.901), all strongly and positively correlated. According to previous studies, these elements mainly originate from vehicular emissions (Chen et al., 2023), tire wear, and lubricant residues (Hou et al., 2019). In particular, vehicular exhaust is a major source of Cd in Beijing (Liu et al., 2007) Therefore, PC1 in ground dust can be attributed to traffic-related sources.
In rooftop dust, Zn (0.913) and Pb (0.942) also show high loadings, while Cd shows a lower loading (0.407), suggesting that Cd at rooftop levels may mainly derive from secondary sources such as building material erosion (Liu et al., 2007). Consequently, PC2 in rooftop dust (with a variance contribution of 30.59%) likely represents traffic emissions.
PC2 in ground dust is dominated by V (0.855) and As (0.959), with strong correlations, reflecting a source related to fossil fuel combustion (Wang, et al., 2023). In rooftop samples, combustion-related elements are mainly concentrated in PC1 (35.79% contribution), with high loadings for V (0.811), Ni (0.856), and As (0.905). The presence of Ni, a known tracer for combustion processes (Turap et al., 2019), suggests that rooftop dust more effectively captures regional atmospheric emissions.
In ground dust, PC3 is characterized by high loadings of Cr (0.855) and moderate loadings of Ni (0.539). The concentrations of both elements are close to Beijing’s background values, indicating a likely origin from natural sources, such as geological background or soil parent materials (Zhang et al., 2021).
For rooftop dust, PC3 is mainly represented by Cu (0.786). Research indicates that Cu is a significant component of construction and industrial activities (Wang et al., 2019), and fine Cu-containing particles can be released into the environment through wind-driven dispersion, accumulating on rooftop surfaces (Men et al., 2020). Additionally, Cu may also derive from wear of vehicle components (Li et al., 2024), suggesting that PC3 in rooftop dust reflects a mixed source of traffic and industrial emissions.
In summary, while ground and rooftop dust share some similarities in heavy metal sources, there are clear differences in the principal component loading structures and variance contributions. This reflects vertical heterogeneity in pollution sources: ground dust is more influenced by localized traffic emissions, whereas rooftop dust better captures regional combustion and construction-related sources.
Element | Rotated Component Matrix (Ground Dust) | Rotated Component Matrix (Rooftop Dust) | ||||
|---|---|---|---|---|---|---|
PC1 | PC2 | PC3 | PC1 | PC2 | PC3 | |
V | 0.012 | 0.855 | 0.392 | 0.811 | -0.005 | 0.209 |
Cr | 0.4 | 0.317 | 0.855 | 0.521 | 0.642 | 0.354 |
Ni | 0.373 | 0.677 | 0.539 | 0.856 | 0.258 | |
Cu | 0.618 | 0.72 | 0.143 | 0.362 | 0.231 | 0.786 |
Zn | 0.953 | 0.032 | 0.23 | 0.205 | 0.913 | 0.005 |
As | 0.238 | 0.959 | 0.089 | 0.905 | 0.172 | -0.043 |
Cd | 0.908 | 0.267 | 0.203 | 0.457 | 0.407 | -0.719 |
Pb | 0.901 | 0.304 | 0.233 | 0.013 | 0.942 | -0.054 |
Variance% | 41.026 | 36.15 | 16.912 | 35.793 | 30.594 | 16.352 |
Cumulative Variance% | 41.026 | 77.176 | 94.087 | 35.793 | 66.387 | 82.739 |
APCS-MLR
The APCS–MLR model was applied to identify the potential sources of heavy metals in dust samples collected at different heights, and the results are presented in Fig. 5.
For ground dust, four types of pollution sources were identified. The fuel combustion source was dominant, accounting for 42.4% of the total contribution. This source was characterized by high levels of As, Cu, and Ni, indicating the influence of coal, wood, and other combustion activities. The traffic source contributed 18.2%, with Zn, Cd, and Pb as dominant elements, suggesting emissions from vehicular exhaust, tire wear, and brake abrasion as the main contributors. The natural source had a minor contribution of only 1.5%, likely associated with natural wind erosion or resuspended soil dust. Additionally, 37.9% of the heavy metal content could not be clearly attributed to known sources and was thus classified as unknown source. This group showed elevated levels of Cr, V, and Ni, possibly linked to complex industrial emissions or long-range atmospheric transport.
Compared to the ground, the source composition of rooftop dust differed significantly. The unknown source was the most dominant, with a contribution of 67.4%, suggesting that rooftop dust is more influenced by regional atmospheric transport or mixed pollution sources. The fuel combustion and traffic sources accounted for only 16.0% and 10.9%, respectively—substantially lower than in ground dust. This reduction is likely due to the rooftop’s relative distance from direct emission sources, resulting in less influence from localized pollutants.
In addition, a traffic–industrial mixed source was identified, contributing 5.7%. This source was characterized primarily by Cu and Cd, indicating a possible combined influence of vehicular activities and industrial emissions.
Overall, these findings highlight significant spatial variability in the sources of heavy metals at different sampling heights. Ground dust better reflects the characteristics of localized anthropogenic activities, while rooftop dust is more affected by regional-scale transport, long-range inputs, and external environmental conditions, making source identification more complex.
Given that the sources of heavy metals are often influenced by multiple overlapping factors (Chen et al., 2022), and that many elements are shared across different sources, the results indicate a multiplicity of sources. Therefore, accurately identifying the key sources and associated indicator metals is critical for effective pollution control and for mitigating health risks (Wang et al., 2021).
Fig. 5
Contribution rate and average contribution rate of heavy metal pollution sources with different heights of dust accumulation (Top: Ground, Bottom: Rooftop)
Conclusions
The average concentrations of Cu, Zn, Cd, and Pb in ground dust in the study area all exceed the background values for Beijing soils, being 1.71, 2.32, 6.98, and 1.80 times higher, respectively. In rooftop dust, the average concentrations of Cr, Ni, Cu, Zn, As, Cd, and Pb are 1.10, 1.09, 2.77, 3.28, 1.09, 9.43, and 2.35 times the background levels, respectively. These findings indicate that most heavy metals have begun to show signs of enrichment, with Cd being the most severe contaminant in both ground and rooftop dust, and thus the primary pollution factor.
The mean potential ecological risk index (Erⁱ) values for Cd in ground and rooftop dust are 210.92 and 283.03, respectively, confirming that Cd is the dominant contributor to ecological risk at both levels. Compared to other elements, Cd contributes the most to the comprehensive risk index (RI).
The Pollution Load Index (PLI) ranges from 0.54 to 2.20 in ground dust and from 1.69 to 2.30 in rooftop dust. Ground dust samples are generally classified as slightly polluted, whereas rooftop dust samples are equally divided between slightly and moderately polluted. With the exception of a few sampling sites where the ground PLI is slightly higher, the average PLI for rooftop dust is consistently higher than that for ground dust, indicating greater pollution levels at higher elevations.
By combining Principal Component Analysis (PCA) with the APCS–MLR receptor model, the source apportionment of heavy metals reveals that in ground dust, the sources are: Fuel combustion (42.4%) > Unknown sources (37.9%) > Traffic emissions (18.2%) > Natural sources (1.5%). In contrast, rooftop dust is mainly influenced by:Unknown sources (67.4%) > Fuel combustion (16.0%) > Traffic emissions (10.9%) > Traffic–industrial mixed sources (5.7%)
A
Author Contribution
Chen Su wrote the main manuscript text , Wenji Zhao contributed to the concepts and ideas of the article, Jie Dong focused on data collection for the article, and Yang Zhiqiang and Zhong Yixue conducted the survey and software.All authors reviewed the manuscript.
References
A
Ali-Taleshi, M.S., Feiznia, S., Masiol, M., 2022. Seasonal and spatial variations of atmospheric depositions-bound elements over Tehran megacity, Iran: Pollution levels, PMF-based source apportionment and risks assessment. Urban Climate 42, 101113.https://doi.org/10.1016/j.uclim.2022.101113Chen, C.-F., Lim, Y.C., Ju, Y.-R., Albarico, F.P.J.B., Chen, C.-W., Dong, C.-D., 2022. Comparing the applicability of ecological risk indices of metals based on PCA-APCS-MLR receptor models for ports surface sediments. Mar Pollut Bull 185, 114361.https://doi.org/10.1016/j.marpolbul.2022.114361
Chen, M., Wang, L.L., Cao, L., Li, M.Y., Shen, Z.M., 2023. Pollution Assessment and Source Analysis of Heavy Metals in Atmospheric Deposition in a Lead-zinc Smelting City Based on PMF Model. Environmental Science 44, 3450–3462.https://doi.org/10.13227/j.hjkx.202208046
Chen, T.B., Zheng, Y.M., Chen, H., Zheng, G.D., 2004. Background Concentrations of Soil Heavy Metals in Beijing. Environmental Science 117–122.https://doi.org/10.13227/j.hjkx.2004.01.026
Cheng, G., Wang, J.L., Deng, B., 2013. Analysis and Research on Traffic Operation Condition of Expressway–Taking Beijing West Third Ring Road as an Example. Highways & Automotive Applications 62–64.
Duan, H.J., Ma, J.Y., Peng, C.Y., Liu, D.X., Wang, Y.F., Liu, D.X., Ma, J.H., 2023. Pollution Characteristics and Source Analysis of Heavy Metals in Soils in Yellow River Cultural Park Based on APCS-MLR and PMF Receptor Model. Environmental Science 44, 4406–4415.https://doi.org/10.13227/j.hjkx.202208173
Duan, H.J., Shen, H.X., Peng, C.Y., Ren, C., Wang, Y.F., Liu, D.X., Wang, Y.L., Guo, R.C., Ma, J.H., 2024. Source Apportionment and Health Risk Assessment of Heavy Metals in Dust Around Bus Stops in Kaifeng City Based on APCS-MLR Model. Environmental Science 45, 3502–3511.https://doi.org/10.13227/j.hjkx.202307013
Fan, X., Lu, X., Yu, B., Zuo, L., Fan, P., Yang, Y., Zhuang, S., Liu, H., Qin, Q., 2021. Risk and sources of heavy metals and metalloids in dust from university campuses: A case study of Xi’an, China. Environmental Research 202, 111703.https://doi.org/10.1016/j.envres.2021.111703
Hakanson, L., 1980. An ecological risk index for aquatic pollution control.a sedimentological approach. Water Research 14, 975–1001.https://doi.org/10.1016/0043-1354(80)90143-8
Hou, S., Zheng, N., Tang, L., Ji, X., Li, Y., Hua, X., 2019. Pollution characteristics, sources, and health risk assessment of human exposure to Cu, Zn, Cd and Pb pollution in urban street dust across China between 2009 and 2018. Environment International 128, 430–437.https://doi.org/10.1016/j.envint.2019.04.046
A
Jo, W.-K., Lee, J.-Y., 2006. Indoor and outdoor levels of respirable particulates (PM10) and Carbon Monoxide (CO) in high-rise apartment buildings. Atmospheric Environment 40, 6067–6076.https://doi.org/10.1016/j.atmosenv.2006.05.037A
Li, D.Z., Guan, Y.T., Liu, A., Li, S.Y., 2015. Speciation Distribution and Risk Assessment of Heavy Metals in Typical Material Roof Dusts. Environmental Science 36, 3269–3277.https://doi.org/10.13227/j.hjkx.2015.09.019Li, J., Li, X., Li, K.M., Jiao, L., Tai, X.S., Zang, F., Cao, S.Z., 2024. Characteristics and Identification Priority Source of Heavy Metals Pollution in Farmland Soils in the Yellow River Basin. Environmental Science 45, 1724–1738.https://doi.org/10.13227/j.hjkx.202305211
Li, J., Teng, Y.G., Wu, J., Jiang, J.Y., Huang, Y., 2019. Source Apportionment of Soil Heavy Metal in the Middle and Upper Reaches of Le'an River based on PMF Model and Geostatistics. Research of Environmental Sciences 32, 984–992.https://doi.org/10.13198/j.issn.1001-6929.2018.07.14
A
Li, X.Y., Zhang, S.T., 2015. Seasonal Provincial Characteristics of Vertical Distribution of Dust Loadings and Heavy Metals near Surface in City. Environmental Science 36, 2274–2282.https://doi.org/10.13227/j.hjkx.2015.06.048Li, Y., Chen, H., Teng, Y., 2020. Source apportionment and source-oriented risk assessment of heavy metals in the sediments of an urban river-lake system. Science of The Total Environment 737, 140310.https://doi.org/10.1016/j.scitotenv.2020.140310
Li, Y.M., Ma, J.H., Liu, D.X., Sun, Y.L., Chen, Y.F., 2015. Assessment of Heavy Metal Pollution and Potential Ecological Risks of Urban Soils in Kaifeng City, China. Environmental Science 36, 1037–1044.https://doi.org/10.13227/j.hjkx.2015.03.037
Liu, C.H., Cen, K., 2007. Chemical composition and possible sources of elements in street dusts in Beijing. Acta Scientiae Circumstantiae 07, 1181–1188.https://doi.org/10.13671/j.hjkxxb.2007.07.018
Liu, L., Guan, D., Yu, Y., Wang, G., 2012. The characteristics of particulate size and heavy metal concentration in foliar dust and surface dust among different urban areas in Guangzhou, China | Request PDF. Fresenius environmental bulletin 21, 1522–1532.
Liu, Z., Du, Q., Guan, Q., Luo, H., Shan, Y., Shao, W., 2023. A Monte Carlo simulation-based health risk assessment of heavy metals in soils of an oasis agricultural region in northwest China. Science of The Total Environment 857, 159543.https://doi.org/10.1016/j.scitotenv.2022.159543
Lu, S.Q., Zhao, Y.N., Xin, Z.M., Duan, R.B., Xiao, H.J., 2020. Comparative Study on Different Coal Dust Suppression Measures in Coal Yard of Mining Area. Journal of Ecology and Rural Environment 36, 945–952.https://doi.org/10.19741/j.issn.1673-4831.2019.0781
Lu, X., Wu, X., Wang, Y., Chen, H., Gao, P., Fu, Y., 2014. Risk assessment of toxic metals in street dust from a medium-sized industrial city of China. Ecotoxicology and Environmental Safety 106, 154–163.https://doi.org/10.1016/j.ecoenv.2014.04.022
Lu, X.H., Yu, F.Z., Fan, Y.M., Yang, Y., 2023. Spatial Distribution Characteristics and Source Analysis of Heavy Metals in Soil of a Lead Plant in Sanmenxia. Environmental Science 44, 1646–1656.https://doi.org/10.13227/j.hjkx.202205052
Ma, Z.Q., Wang, Y.S., Sun, Y., Ji, D.S., 2007. Characteristics of Vertical Air Pollutants in Beijing. Research of Environmental Sciences 1–6.https://doi.org/10.13198/j.res.2007.05.3.mazhq.001
Men, C., Liu, R., Xu, L., Wang, Q., Guo, L., Miao, Y., Shen, Z., 2020. Source-specific ecological risk analysis and critical source identification of heavy metals in road dust in Beijing, China. Journal of Hazardous Materials 388, 121763.https://doi.org/10.1016/j.jhazmat.2019.121763
Ministry of Ecology and Environment of the People’s Republic of China. Solid Waste–Determination of metals–Inductively coupled plasma mass spectrometry (ICP-MS) (HJ 766–2015).https://www.mee.gov.cn/ywgz/fgbz/bz/bzwb/jcffbz/201511/t20151130_317999.shtml
Mummullage, S., Egodawatta, P., Ayoko, G.A., Goonetilleke, A., 2016. Use of physicochemical signatures to assess the sources of metals in urban road dust. Science of The Total Environment 541, 1303–1309.https://doi.org/10.1016/j.scitotenv.2015.10.032
Ramírez, O., Sánchez de la Campa, A.M., Sánchez-Rodas, D., de la Rosa, J.D., 2020. Hazardous trace elements in thoracic fraction of airborne particulate matter: Assessment of temporal variations, sources, and health risks in a megacity. Science of The Total Environment 710, 136344.https://doi.org/10.1016/j.scitotenv.2019.136344
Roy, S., Gupta, S.K., Prakash, J., Habib, G., Kumar, P., 2022a. A global perspective of the current state of heavy metal contamination in road dust. Environ Sci Pollut Res Int 29, 33230–33251.https://doi.org/10.1007/s11356-022-18583-7
Shi, D.Q., Lu, X.W., 2018. Contamination Levels and Source Analysis of Heavy Metals in the Finer Particles of Urban Road Dust from Xi'an, China. Environmental Science 39, 3126–3133.https://doi.org/10.13227/j.hjkx.201711072
Shi, W., Li, T., Feng, Y., Su, H., Yang, Q., 2022. Source apportionment and risk assessment for available occurrence forms of heavy metals in Dongdahe Wetland sediments, southwest of China. Science of The Total Environment 815, 152837.https://doi.org/10.1016/j.scitotenv.2021.152837
Song, Y., Xie, S., Zhang, Y., Zeng, L., Salmon, L.G., Zheng, M., 2006. Source apportionment of PM2.5 in Beijing using principal component analysis/absolute principal component scores and UNMIX. Science of The Total Environment 372, 278–286.https://doi.org/10.1016/j.scitotenv.2006.08.041
Thurston, G.D., Spengler, J.D., 1985. A quantitative assessment of source contributions to inhalable particulate matter pollution in metropolitan Boston. Atmospheric Environment (1967) 19, 9–25.https://doi.org/10.1016/0004-6981(85)90132-5
Tomlinson, D.L., Wilson, J.G., Harris, C.R., Jeffrey, D.W., 1980. Problems in the assessment of heavy-metal levels in estuaries and the formation of a pollution index. Helgolander Meeresunters 33, 566–575.https://doi.org/10.1007/BF02414780
Turap, Y., Talifu, D., Wang, X., Abulizi, A., Maihemuti, M., Tursun, Y., Ding, X., Aierken, T., Rekefu, S., 2019. Temporal distribution and source apportionment of PM2.5 chemical composition in Xinjiang, NW-China. Atmospheric Research 218, 257–268.https://doi.org/10.1016/j.atmosres.2018.12.010
Wang, H.-L., Shen, C.-Y., Li, P.-L., Faisal, M., Zhu, S., 2021. Vertical Distribution Characteristics and Source Apportionment Of Heavy Metals in Urban Near-Surface Dust Based on Receptor Model. Polish Journal of Environmental Studies 30, 5811–5831.https://doi.org/10.15244/pjoes/137665
Wang, L., Lu, X., Li, L.Y., Ren, C., Luo, D., Chen, J., 2015. Content, speciation and pollution assessment of Cu, Pb and Zn in soil around the lead–zinc smelting plant of Baoji, NW China. Environ Earth Sci 73, 5281–5288.https://doi.org/10.1007/s12665-014-3777-5
Wang, S., Cai, L.-M., Wen, H.-H., Luo, J., Wang, Q.-S., Liu, X., 2019. Spatial distribution and source apportionment of heavy metals in soil from a typical county-level city of Guangdong Province, China. Science of The Total Environment 655, 92–101.https://doi.org/10.1016/j.scitotenv.2018.11.244
Wang, X.Q., Fei, X.H., Yang, Y., Li, Y.W., Gui, J.Q., Yang, A.J., Xu, P., 2023. Pollution characteristics, source apportionment and health risk assessment of heavy metal elements in PM2.5collected in Huaxi urban areas, Guiyang. Acta Scientiae Circumstantiae 43, 110–118.https://doi.org/10.13671/j.hjkxxb.2022.0473
Wilding, L.P., Wilding, L.P., 1985. Spatial variability: its documentation, accommodation and implication to soil surveys. Soil spatial variability 166–189.
A
Yadav, I.C., Devi, N.L., Singh, V.K., Li, J., Zhang, G., 2019. Spatial distribution, source analysis, and health risk assessment of heavy metals contamination in house dust and surface soil from four major cities of Nepal. Chemosphere 218, 1100–1113.https://doi.org/10.1016/j.chemosphere.2018.11.202Zhang, W., Yan, Y., Yu, R., Hu, G., 2021. The sources-specific health risk assessment combined with APCS/MLR model for heavy metals in tea garden soils from south Fujian Province, China. CATENA 203, 105306.https://doi.org/10.1016/j.catena.2021.105306
Zhang, Y., Zhang, Y.F., Hu, X.,2010. Assessment of Heavy Metal Contamination and Source Identification of Dust Deposited on Roads Collected from Different Land-Use Zones in Nanjing. Research of Environmental Sciences 23, 1376–1381.https://doi.org/10.13198/j.res.2010.11.47.zhangy.006
Zhao, G., Zhang, R., Han, Y., Meng, J., Qiao, Q., Li, H., 2021. Pollution characteristics, spatial distribution, and source identification of heavy metals in road dust in a central eastern city in China: a comprehensive survey. Environ Monit Assess 193, 796.https://doi.org/10.1007/s10661-021-09584-z
Zhao, H., Huang, T., Su, J., Li, X., 2023. Role of height and position in the vertical distribution pattern of urban surface-deposited sediments and associated heavy metals. Environ. Sci.: Water Res. Technol. 9, 3146–3157.https://doi.org/10.1039/D2EW00813K
Zou, C.L., Zhao, Q.S., Xie, W.X., Li, X.Y., Zhao, H.T., 2017. Effects of depositional height on the distribution and heavy metal content of urban dust. Acta Scientiae Circumstantiae 37, 729–736.https://doi.org/10.13671/j.hjkxxb.2016.0298