To our best knowledge, this study is the first to examine the source-specific ecological risk of PM2.5-bound metals in a multi-city region over a relatively long period of time (i.e., > 1.5 years), and also the first to evaluate the ecological risk of PM2.5-bound metals and sources in the PRD region, China. The PERIs of the selected PM2.5-bound metals all exceeded 600, suggesting a high potential ecological risk in this region. The potential ecological risk of PM2.5-bound metals and sources in the PRD showed seasonal variation and spatial patterning. Higher concentrations of PM2.5-bound metals occurred in the dry season than in the wet season, and in the northwest when compared with the other two clusters. The PM2.5-bound metals from industrial emissions and vehicle emissions were estimated to contribute most to the ecological risk in the PRD region, followed by sources of coal burning, dust emissions and heavy oil burning.
Spatial-temporal characteristics of ecological risk of PM 2.5 -bound metals. According to our comprehensive evaluation, an extreme ecological risk (PERI > 600) was widespread in the PRD region. These high potential ecological risks are mainly caused by the toxic-response value and the concentration of Cd. This finding is consistent with the previous ecological risk assessment results reported for PM2.5-bound metals based on a single city or a single sampling site (Ghasemi et al. 2020; Zhai et al. 2014). PM2.5 pollution has been a long-standing concern in densely populated and industrialized areas in China, and the PRD was no exception before the year 2013. Duan et al. (2013) found that major cities in the northwestern city cluster of PRD (e.g., Foshan) were ranked among the most atmospheric heavy metal polluted cities in China, with atmospheric Cd concentration 12 times higher than the World Health Organization limit [5 ng/m3] (WHO, 2000) in 2010. Problems with a PM2.5 pollution in this region have been mitigated since the China carried out the national Air Pollution Prevention and Control Action Plan (APPCAP) with a series of emission control policies in 2013. The concentration of atmospheric Cd of 2015–2016 observed in our study (1.17 ± 1.48 ng/m3) has dropped by 90% over that of 2010 reported in the Duan et al.’s study. However, even as atmospheric Cd concentration has dramatically declined in recent decade and fallen below the WHO limit, it still can pose extremely serious ecological risk to the PRD region.
We found pronounced heterogeneity in regional distribution of the airborne metals’ ecological risk. The PERI in the northwestern city cluster surpassed that in the southwestern or eastern city clusters. This spatial discrepancy is mainly attributed to the effects of the emissions and meteorological processes (Wu et al. 2013; Yuval et al. 2020). For example, the northwestern city cluster, located inland alongside mountain with a high frequency of temperature inversion, is characterized by dynamic industrial activities that entail substantial discharge intensities and produce numerous pollutants. Although no greater industrial activities currently occur in the southwestern city cluster, it is nonetheless located in a perennial downwind area and receives pollutants from both the northwestern and eastern industrial regions, which, to a certain extent, exacerbate metal enrichment. Eastern city clusters are located in coastal areas, where strong sea winds enable pollutants’ diffusion and mitigate enrichment of their bound metals.
We found that PM2.5-bound metals’ ecological risk was more severe in the dry than wet season. This result agrees with other studies in the PRD that reported levels of atmospheric particulate matter pollution (Hu et al. 2021; Lin et al. 2019; Tian et al. 2021). Frequent precipitation during the wet season in the PRD and a thickening of its atmospheric boundary layer due to strong solar radiation are conducive to air pollutant diffusion. At the same time, clean air transport from the ocean maintained air pollutants at lower levels (Liao et al. 2020). By contrast, the meteorological conditions of PRD’s dry season, namely intense solar radiation with less cloud cover and low wind speeds, are unfavourable for atmospheric diffusion and deposition; hence, a vertical column of inverted temperature is more likely to form then aggravating pollution (Han et al. 2014). Source-specific ecological risks also showed distinctive temporal heterogeneity between the dry and wet season, indicating that a well-directed air pollution control strategy tailored to each season should be implemented.
Source-specific ecological risk of PM 2.5 -bound metals. Multiple sources hinder the effectiveness of control strategies and hazard prevention targeting PM2.5 concentrations, the identification of the most responsible sources is critical to facilitate more focused regulatory approaches for PM2.5. In our study, vehicle emissions and industrial emissions were paramount sources of PM2.5-bound metals related to ecological environmental damage across the PRD region. We found that Cd made a disproportionally large contribution to both sources, accounting for 34% of vehicle emissions and 63% industrial emissions respectively (see Supplementary Information, Fig. S3). Among the studied PM2.5-bound metals, those with higher toxic response factor, notably Cd, may responsible for more ecological risk than others. The extremely higher ecological risk related to both vehicle and industrial emissions could therefore closely tied to the toxicity of Cd (Men et al. 2020). Against this region’s background of rapid economic development, it would be prudent to pay more attention to critical pollution sources based on corresponding ecological risks.
The ecological risk of industrial emission sources is dominant in the northwestern urban agglomeration, likely due to intensive industrial activity there and an urban heat island effect. The annual PM2.5 emission level in that northwestern cluster was twice that of the eastern cluster (Huang et al. 2018). Several cities in the PRD have the highest density of motor vehicles in China, leading to a very high ecological risk from vehicle emissions, whether in the dry or wet season. In 2014, the number of motor vehicles reached 13.26 million in Guangdong province, among which PRD was the top one, accounting for more than 80% of the motor vehicle ownership (Wang et al. 2016). In the past few years, in order to reduce exhaust emissions from traffic, several control measures such as eliminating yellow-label vehicles (that is, vehicles with emission level below China I emission standard), tightening emission standards and improving fuel quality have been carried out in the PRD region (Liu et al. 2017). Through these efforts of the local government in the PRD, particulate pollution caused by vehicle exhaust has declined markedly, which the estimated reduction rate was more than 50% (Cheng et al. 2014). However, given the magnitude of the observed ecological risk of non-exhaust emissions of vehicles (e.g., braking and tire wearing) in our study, continued efforts to identify strategies for reducing non-exhaust particulate emissions from road traffic is warranted.
Study limitations and strengths. As conveyed in Table 2, past studies of PM2.5-bound metals have focused on determining the ecological risks of critical metals, and explored their possible sources via well-established methods, but they did not quantify the respective ecological risk of each source. The major strengths of this study are that by collecting long-term in situ multi-site monitoring data, we were well positioned to elaborate on the magnitude of ecological risks stemming from six possible sources of PM2.5-bound metals in the atmosphere, and applied PMF to partition the ecological risks associated with the six metals.
Table 2
Comparison of source-specific ecological risk for the PM2.5-bound metals (ng/m3) in this study and other similar studies.
Previous study | PM2.5-bound metals | Sampling effort | Identified sources (method) | Quantitative (yes, no) |
Zhi et al.( 2021) | 23 metal elements | One site, from January to March 2016 | Anthropogenic sources (EF) | No |
Zhai et al.( 2014) | Cu, Zn, Pb, Cd | Three sites in one city, March to April of 2013 | Traffic sources, fossil fuel, and dust (PCA) | No |
Zhang et al.( 2021) | 12 metal elements | One site, from May 2013 to February 2019 | Ship emissions and traffic sources (EF + CWT) | No |
Ghasemi et al.( 2020) | Cd, Cr, Co, Pb, Fe, V and Ni | One site, from December 2016 to September 2017 | Industrial emission, traffic sources, and latex color additives (PCA) | No |
Present study | Cd, Cr, Cu, Zn, Pb, and As (23 metal elements were measured and used in source apportionment) | Ten sites, from December 2014 to July 2016 | Dust emissions, vehicle emissions, industrial emissions, coal burning, and heavy oil burning (PMF) | Yes |
Abbreviations: EF: enrichment factor; PCA: principal component analysis; CWT: concentration-weighted trajectory analysis; PMF: positive matrix factorization. |
Our study has some limitations. First, to ensure a robust source apportionment, a total of 23 metals were used in the PMF model, but due to limited ecological risk assessment model standards, our study focused on six metals (Cr, Cd, Cu, Zn, As and Pb) for actual ecological risk assessment. This likely underestimated the true ecological risk from PM2.5-bound metal sources, as in any ecological risk assessment study. In future work, better accuracy of ecological risk assessment can be gained by including more metals, and even PM2.5 bounded water-soluble ions and organic species. Secondly, particle sources in fact contain a mixture of pollutants more than the studied metals alone; perhaps some ecological risk could also arise from other unmeasured chemical compositions that were co-generated from the same source or share similar processing. Although the ecological risk in our study may be underestimated to some extent, it is still of great significance to have reliably identified the most harmful sources in the PRD region.