Seasonal Variations and Health Risk Evaluation of Trace Elements in Atmospheric PM_2.5 in Liaocheng, the North China Plain

Abstract

Atmospheric elements can cause harmful effects on air quality and human health. Despite extensive research on PM_2.5, there remains a limited understanding of the seasonal variations, origins, and associated health risks of specific elements in urban areas of the North China Plain. PM_2.5 samples across four seasons were collected to investigate the seasonal variations, provenance, and health risks of 18 elements in urban Liaocheng. The concentrations of PM_2.5 and total detected elements (TDEs) exhibited distinct seasonal patterns, with the biggest values occurring in winter, followed by spring, autumn, and summer. Fe, Ca, Al, and K were the most plentiful elements throughout the campaign, contributing 72.2% of TDEs. The enhanced concentrations of crustal elements were due to frequent dust storms in spring. Results from positive matrix factorization suggested that the dust source was only identified in spring, accounting for the largest percentage (37.0%), while secondary oxidation made the most significant contribution (34.6%) in summer, facilitated by higher temperatures and stronger sunshine. The relative abundance (41.6%) of biomass burning was highest in autumn, ascribed to intensified agricultural waste burning during the autumn harvest, especially in October. The contribution of coal combustion in cold seasons was substantially greater than in warm seasons, highlighting the role of increased coal burning for house heating in deteriorating air quality. Potential source function analysis showed that elements in Liaocheng originated from local and neighboring regions. The carcinogenic risk from the selected elements was notably stronger for adult males than for adult females and children, while the non-carcinogenic risk was stronger for children than for adults. Overall, these findings provide ponderable insights into the contamination characteristics and sources of elements, which are useful to inform effective measures for improving air quality and aerosol modeling.

1. Introduction

Elements are crucial contributors to fine particulate matter (PM_2.5) and influence haze occurrence, atmospheric chemistry, and climate variation [1,2,3]. Atmospheric elements are widely present and pose serious risks to human health through inhalation, skin exposure, and ingestion of pollutants [4,5,6]. In particular, heavy metals have received more concern because of their damage to the respiratory, nervous, and cardiovascular systems [7]. Elements such as Sb, Ni, Cr(VI), Cd, Pb, As, and their compounds have been classified as carcinogenic by the International Agency for Research on Cancer of the World Health Organization [8]. In view of the Agency for Toxic Substances and Disease Registry (ATSDR) [9], Ni and Co can harm the respiratory system; Pb can impair children’s intelligence and harm fetal development; Cd and As can cause human malformation. Additionally, transition elements can serve as catalysts in aqueous oxidation reactions and the heterogeneous generation of secondary organic aerosols, and facilitate the generation of SO₄²⁻ in the particulate phase, thereby promoting haze formation [10]. Specific elements also can be applied as fingerprinting indicators for the source analysis of PM_2.5 [11]. For example, Se and As have been proposed as significant indicators for coal burning [12,13]; Fe can be used as an excellent marker of iron and steel manufacturing [4]. Therefore, atmospheric elements have gained increasing concern recently [14,15,16].

China has become one of the leading contributors of trace elements throughout the world, largely owing to rapid industrial development and increased energy demands [4,17]. The emissions of Cd, As, Pb, and Cr in China rank among the highest globally [18]. Hence, a comprehensive investigation of the pollution characteristics, sources, and health risks of these elements is sorely needed in China, especially in cities experiencing severe air pollution. Numerous studies have measured trace elements across different environments such as urban, suburban, marine, and mountainous regions [19,20,21,22]. Element concentrations throughout China exhibit distinct spatial patterns, exhibiting higher levels in northern China than in southern China, especially for crustal elements [14]. Most elements show elevated concentrations in urban regions compared to rural regions [23,24,25,26]. Liu et al. [16] researched the seasonal patterns of heavy metals in urban Beijing, suggesting that enhanced total heavy metals during the heating season compared to that during the non-heating season. However, most studies on urban aerosols have focused on megacities such as Guangzhou, Shanghai, and Tianjin, with limited attention given to trace elements in smaller cities [27,28,29]. Accordingly, further investigation into the seasonal changes and provenance of elements in small cities with severe air pollution is essential to enhance our understanding of spatial variations across China and to develop effective control measures for the emission of trace elements into the atmosphere.

As one of the “2 + 26” atmospheric transmission channel cities in the Beijing-Tianjin-Hebei (BTH) region of China, Liaocheng is facing severe atmospheric pollution. In 2021, Liaocheng was ranked as the top 21st city with the poorest air quality among 339 cities in China, leading to its designation as a priority city for air quality control by the Ministry of Ecology and Environment of the People’s Republic of China. Many studies have suggested that atmospheric pollutants released from Liaocheng can significantly influence the BTH region, identifying it as a crucial potential origin area for both the BTH region and the North China Plain (NCP) [30,31,32]. Given its unique geographical location, a deep understanding of the pollution characteristics and provenance of elements in Liaocheng is urgently needed to address atmospheric pollution in the NCP. While numerous investigations have been performed in Liaocheng, they have primarily focused on organic aerosols (OAs) and single particles [30,33,34,35]. For instance, Li et al. [30] examined the seasonal changes in OAs in Liaocheng and found that OAs throughout the year are derived from vehicle exhaust, with an enhanced contribution of secondary oxidation to OAs in summer. To our knowledge, the pollution characteristics and health risk evaluation of trace elements in Liaocheng have not been reported, especially across the four seasons. This work aimed to compare the molecular distributions, origins, and health risks of elements in the urban aerosols of Liaocheng across four seasons, which can provide crucial theoretical support for air quality management in Liaocheng and the broader NCP region.

2. Materials and Methods

2.1. Sample Collection

Liaocheng City, situated in the central part of the NCP, has been experiencing severe air pollution. It possesses a representative monsoon climate with four distinctive seasons. The sampling campaign was conducted at a national air quality monitoring station located atop a teaching building (36.43° N, 116.01° E; 25 m above ground) at Liaocheng University. The sampling site, surrounded by residential areas and heavily trafficked roads without any industrial emissions or tall buildings, serves as a representative location for evaluating urban air quality in Liaocheng. In our sampling campaign, sampling activities extended for 1 month in each season to guarantee the consistency of the sampling period across seasons. In spring, the traditional coal-fired heating period is routinely stopped on 15 March 2021 in Liaocheng. Due to the sudden reduction in temperature from 14 March to 22 March in 2021, the heating period was extended to 23 March 2021 in Liaocheng. To avoid the effect of the coal combustion used for heating, the data during the heating period were eliminated in this study. Similarly, we also excluded some samples collected from 5 April to 10 April, when dust storms occurred in Liaocheng, to eliminate the impact of dust storms. During the summertime period, samples collected from 28 July to 3 August, when typhoons occurred, were excluded to eliminate the effect of typhoons on elements. During the autumn sampling, constant rain occurred from 7 October to 15 October; thus, these samples were excluded to eliminate the effect of wet deposition on elements. Therefore, PM_2.5 samples were collected on a seasonal basis in 2021, covering spring (24 March to 4 April), summer (4 July to 27 July), autumn (16 September to 6 October), and winter (1 December to 31 December) to investigate the seasonal characteristics of elements. Quartz fiber filters have the advantage of high-temperature resistance, strong chemic inertia, and low background values; thus, they have been widely applied for the measurement of water-soluble inorganic ions and metals [36,37,38,39,40,41]. In this work, PM_2.5 samples were collected onto preheated (450 °C for 6 h) quartz filters (90 mm diameter, Whatman) using a medium-flow sampler (Model: 2030, Qingdao Laoying Co., Ltd., Qingdao, China), with a velocity of 100 L min⁻¹. Sample collection was divided into daytime (08:00–19:30) and nighttime (20:00–07:30 the next day). In this work, while diurnal changes in elements were not analyzed, daytime and nighttime concentrations were averaged to examine seasonal differences in element concentrations. Field blank samples were also obtained by putting blank filters onto the sampler for 10 min without sucking air before, during, and after sampling, respectively. Six field blank samples were obtained in this study. Finally, all samples were wrapped in aluminum foil paper and stored at −20 °C prior to analysis. Data on gaseous pollutants, including CO, O₃, SO₂, and NO_x, were downloaded from the Shandong Integrated Atmospheric Observation Platform (http://117.73.252.128:9010/saas/login, accessed on 18 December 2024). Meteorological parameters were retrieved from the Liaocheng Meteorological Bureau (http://sd.cma.gov.cn/gslb/lcsqxj/, accessed on 18 December 2024).

2.2. Sample Analysis

The mass concentration of PM_2.5 was determined by weighing the filters with a microbalance (Mettle M3, Switzerland, accuracy of 0.1 μg) described in Zhang et al. [42]. Before weighing, the filters were placed in a constant relative humidity (40 ± 5.0%) and temperature (20 ± 1.0 °C) environment for 48 h. Before and after sampling, all the filters were weighed twice to ensure error within 5.0 μg. An electrostatic elimination device was also used to guarantee the accuracy of weight results. The sampled PM_2.5 mass was calculated by the difference in the filter mass after and before sampling.

In this study, a total of 27 elements, including Ba, Cr, Co, Al, Ca, Fe, Mn, Cu, Ni, Pb, V, As, K, S, Cl, Ti, Mg, Zn, P, Si, Se, Br, Sc, Ga, Ge, Sr, and Bi, were quantitatively determined in collected samples by an energy-dispersive X-ray fluorescence spectrometer (ED-XRF). Among them, nine elements (P, Si, Se, Br, Sc, Ga, Ge, Sr, and Bi) in most PM_2.5 samples were below the detection limits. The concentrations of these nine elements were so low that they had little effect on the total concentrations, origins, and health risks of elements. Therefore, 18 elements, including Ba, Cr, Co, Al, Ca, Fe, Mn, Cu, Ni, Pb, V, As, K, S, Cl, Ti, Mg, and Zn, were discussed in the current work. To optimize accuracy across this range of elements, four distinct analytical conditions were sequentially applied, with each condition tailored to excite specific groups of elements. The X-ray tube, capable of a maximum current of 1660 µA and a maximum voltage of 50 kV, was adjusted according to the requirements for each set, as detailed in Table S1. The initial conditions were used for Mg, Al, S, and Cl; the second set of conditions for K, Ca, Ti, and Ba; the third for V, Cr, and Mn, and the fourth for Fe, Co, As, Pb, Ni, Zn, and Cu. The analytical parameters for each step are outlined in Table S1. Every sample was analyzed in a 20-minute cycle. An autosampler equipped with a turret holding 10 cassettes facilitated the batch analysis of 10 samples without manual replacement. One sample per batch was randomly selected for a repeat measurement to ensure that the analytical error in element concentrations remained less than 2.0%.

For the measurement of NO₃⁻, SO₄²⁻, NH₄⁺, and K⁺, a quarter of each sample was extracted with 10 mL Milli-Q water (R > 18.2 MΩ) using an ultrasonic bath three times, and then filtered through PTFE filters to remove the insoluble substances. The water extract was then measured for NO₃⁻, SO₄²⁻, NH₄⁺, and K⁺ by ion chromatography (Dionex Aquion and Dionex-600, Thermo Fisher Scientific, California, USA). All the concentrations of target chemical species in field blank samples were lower than 5% of the concentrations of the collected field samples. All the concentrations of 18 elements and NO₃⁻, SO₄²⁻, NH₄⁺, and K⁺ in each field blank sample are summarized in Table S2. The targeted chemical species presented here were corrected for the average concentrations of field blanks.

2.3. Enrichment Factor Analysis

Enrichment factors (EFs) of elements have been commonly used to distinguish anthropogenic and crustal sources [43]. They were calculated using the following formula:

EFs = (C_Element/C_Refercence)_Aerosol/(C_Element/C_Refercence)_Crust,

(1)

where C_Element represents the concentration of a specific element, and C_Reference represents the concentration of a reference element. The subscript “Aerosol” refers to element concentrations in aerosol samples, while “Crust” denotes element concentrations in the Earth’s upper crust. Fe, Al, and Si, abundant and stable in the crust, are commonly chosen as reference elements [44]. Al was selected as the reference element because of its high chemical stability [45]. Data for crustal element concentrations were sourced from Taylor and McLennan [46]. EF values close to 1 indicate that elements primarily originate from natural sources with minimum enrichment; values between 1 and 10 suggest mixed sources, with natural sources dominating and slight enrichment from anthropogenic activities; and values ranging from 10 to 100 reflect a more substantial impact of anthropogenic emissions with moderate enrichment, while values exceeding 100 denote a dominance of anthropogenic sources with high enrichment [47,48].

2.4. Positive Matrix Factorization (PMF) Modeling

The PMF 5.0 model developed by the Environmental Protection Agency has been widely used to identify sources of atmospheric elements [49]. The model decomposes the sample data matrix into a factor contribution matrix and factor profile matrix, allowing the calculation of source types and their respective contributions based on input data. The PMF model is described as follows:

$${{\displaystyle X}}_{ij}={\displaystyle {\displaystyle \sum }_{k=1}^P{{\displaystyle g}}_{ik}{{\displaystyle f}}_{ki}+{{\displaystyle e}}_{ij,}}$$

(2)

where X_ij represents the concentration of species j detected in sample i, p is the number of contribution factors, g_ik is the relative contribution of factor k to sample i, f_kj is the mass contribution of species j in factor k, and e_ij is the error of species j in sample i. To obtain the optimal matrices g and f, the matrix x is decomposed to minimize the objective function Q, defined as

$$Q={\displaystyle \sum _{i=1}^{\mathrm{n}}{\displaystyle \sum _{j=1}^m(\frac{{{\displaystyle x}}_{ij}-{\displaystyle \sum _{k=1}^P{{\displaystyle g}}_{ik}{{\displaystyle f}}_{ki}}}{{{\displaystyle u}}_{ij}})}}={\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m{{\displaystyle (\frac{{{\displaystyle e}}_{ij}}{{{\displaystyle u}}_{ij}})}}^2},}$$

(3)

where u_ij is the uncertainty of the concentration of species j in sample i. If the measured species concentration is below the method detection limit (MDL), it is replaced with 1/2 MDL, and the corresponding uncertainty is set to 5/6 MDL. When the concentration is above the MDL, the uncertainty is calculated as

$$\mathrm{Uncertainty}=\sqrt{{(\mathit{error} \mathit{fraction}\times \mathit{concentration})}^2{+(0.5\times \mathit{MDL})}^2},$$

(4)

where the error fraction represents the estimated analytical uncertainty of the measured concentration or flux [50,51,52].

2.5. Health Risk Evaluation

A health risk evaluation model was applied to assess the toxic effects of specific metals inhaled by children, adult males, and adult females. According to the Inhalation Dosimetry Methodology established by the USEPA [53], health risks associated with both non-carcinogenic (Co, As, Pb, Mn, Zn, Ni, and V) and carcinogenic (Co, As, Ni, and Pb) elements were assessed. The Lifetime Average Daily Dose (LADD) for both non-carcinogenic and carcinogenic metals was calculated as follows [54]:

$$\mathrm{LADD} \& \mathrm{ADD}=(\mathrm{c} \times \mathrm{IR} \times \mathrm{EF} \times \mathrm{ED})/(\mathrm{BW} \times \mathrm{AT}),$$

(5)

where c represents the concentrations of metals (mg m^–3), and the values for IR, ED, EF, BW, and AT for adults and children are as summarized in

Table 1. Parameters used for assessing non-carcinogenic and carcinogenic health risks.

Parameters	Abbreviation	Children	Adult Males	Adult Females
Inhalation rate	IR	8.6 m3 day−1	16.6 m3 day−1	13.5 m3 day−1
Exposure frequency	EF	365 days year−1	365 days year−1	365 days year−1
Exposure duration	ED	6 years	24 years	24 years
Average body weight	BW	15 kg	67.3 kg	57.5 kg
Average time (non-carcinogens)	AT	ED × 365 days	ED × 365 days	ED × 365 days
Average time (carcinogens)	AT	70 × 365 days	70 × 365 days	70 × 365 days

Note: Parameters were cited from He et al. [55] and Wang et al. [56].

.

The health risks of metals were then calculated using the following equation. The cancer risk for each element was defined as

$$\mathrm{CR}=\mathrm{LADD} \times \mathrm{SF},$$

(6)

with the total cancer risk (TCR) being

$$\mathrm{TCR}={\sum \mathrm{CR}}_i,$$

(7)

The non-cancer risk, the hazard quotient (HQ), was calculated as

$$\mathrm{HQ}=\mathrm{ADD}/\mathrm{RfD},$$

(8)

and the hazard index (HI), total cumulative non-carcinogenic risk, was defined as

$$\mathrm{HI}=\sum \mathrm{H}{\mathrm{Q}}_{i,}$$

(9)

where CR represents the cancer risk of an element (unitless value), SF represents the slope factor of an element (kg day mg⁻¹), TCR is the total cancer risk (unitless value), HQ stands for the hazard quotient for assessing the non-carcinogenic risk of an element (unitless value), RfD is the reference dose (mg kg⁻¹ day⁻¹), and HI stands for the hazard index (unitless value). The values of RfD and SF for each selected element are summarized in

Table 2. Reference doses (RfD) and slope factors (SF) of the seven selected toxic elements.

Toxic Elements	RfD (mg (kg day)−1)	SF ((kg·day) mg−1)
As	3.01 × 10−4	1.51 × 101
Ni	2.06 × 10−2	8.40 × 10−1
Co	5.71 × 10−6	9.80 × 10
Pb	3.52 × 10−3	8.50 × 10−3
Mn	1.43 × 10−5	−
Zn	3.00 × 10−1	−
V	7.00 × 10−3	−

Note: Dates were cited from Wang et al. [56] and. Li et al. [57].

.

2.6. Backward Trajectories and Potential Source Contribution Function (PSCF) Simulation

The 48-hour backward trajectories (100 m above the ground) were simulated using the Hybrid Single-particle Lagrangian Integrated Trajectory (HYSPLIT) model (https://www.ready.noaa.gov/HYSPLIT_traj.php, accessed on 18 December 2024) [58]. This model was operated every 6 h, and the results were saved as endpoint files. PSCF was used to determine the potential source regions of elements. The study areas were divided into equal i × j grid cells (ij), and the results were defined as follows:

$$PSCF\mathit{ij} ={\displaystyle \frac{m_{\mathit{ij}}}{n_{\mathit{ij}}}},$$

(10)

where n_ij represents the number of endpoints through the ij grid cell (0.5° × 0.5° in this study). m_ij is the value of trajectory endpoints concentrated in the same cell. The PSCF_ij values were multiplied by an arbitrary weight function W_ij to lower the uncertainties.

$$W_{\mathit{ij}} =\left\{\begin{array}{c}1.00,3{\mathrm{n}}_{ave}<{\mathrm{n}}_{ij}\hfill \\ 0.70,{\mathrm{n}}_{ave}<{\mathrm{n}}_{ij}<3{\mathrm{n}}_{ave}\hfill \\ 0.42,0.5{\mathrm{n}}_{ave}<{\mathrm{n}}_{ij}<{\mathrm{n}}_{ave} \hfill \\ 0.05,{\mathrm{n}}_{ij}<0.5{\mathrm{n}}_{ave},\hfill \end{array}\right.$$

(11)

Here, the PSCF algorithms were calculated by the TrajStat plugins in MeteoInfoMap software(Version: 2.3.5), which has been proven effective in identifying the potential source regions of atmospheric pollutants [58].

In summary, a flowchart of this work is depicted in Figure S1.

3. Results and Discussion

3.1. Seasonal Patterns of PM_2.5 Mass and Elements

The average annual PM_2.5 concentration in Liaocheng throughout the sampling period was 65.9 ± 31.5 μg m⁻³, approximately twice the Grade−II level (35 μg m⁻³) of China’s national ambient air quality standard (NAAQS). This concentration exceeded levels in other Chinese cities such as Beijing (33.0 μg m⁻³) [27,59], Shanghai (30.1 μg m⁻³) [60], Zhengzhou (46.0 μg m⁻³) [19], Shenzhen (17.5 μg m⁻³) [61], and Chongqing (37.5 μg m⁻³) [62] in 2021. These findings indicate that PM_2.5 pollution in Liaocheng warrants significant attention, and implementing control measures to mitigate air pollution is crucial. PM_2.5 concentrations in Liaocheng displayed distinct seasonal variations, with higher concentrations in colder seasons (winter: 89.0 ± 32.2 μg m⁻³; spring: 80.0 ± 28.3 μg m⁻³) than in warmer seasons (autumn: 48.6 ± 18.1 μg m⁻³; summer: 44.3 ± 10.0 μg m⁻³) (

Table 3. Seasonal variations in elements, secondary inorganic ions, meteorological parameters, and PM_2.5 mass during the whole observation period in Liaocheng.

Parameter	Spring	Summer	Autumn	Winter	Annual
Ⅰ. Elements (ng m−3)
Al	795 ± 183	551 ± 28.3	608 ± 74.9	714 ± 35.5	667 ± 172
Ca	587 ± 130	166 ± 49.8	184 ± 64.3	632 ± 81.6	393 ± 189
Fe	372 ± 256	122 ± 21.8	116 ± 23.0	318 ± 60.7	232 ± 235
K	207 ± 258	80 ± 10.3	120 ± 14.6	256 ± 43. 1	166 ± 120
Ti	30.2 ± 44.2	6.5 ± 1.4	12.3 ± 1.4	22.3 ± 4.7	17.8 ± 17.7
Mg	97.9 ± 40.6	51.2 ± 3.0	47.8 ± 4. 1	72.4 ± 8.8	67.3 ± 7.6
∑Crust	2089 ± 507	977 ± 332	1088 ± 317	2015 ± 469	1542 ± 512
S	118 ± 65.2	223 ± 36.6	139 ± 39.9	277 ± 61. 1	189 ± 59.9
Cl	76.5 ± 16.0	34.9 ± 15.2	104 ± 10.3	391 ± 13.3	151 ± 19.7
Zn	67.3 ± 9.3	56.5 ± 3.5	58.4 ± 5.3	108 ± 7.0	72.5 ± 6.9
Mn	38. 1 ± 14.0	9.9 ± 0.7	15.4 ± 1. 1	12.9 ± 2.3	19. 1 ± 2.3
Ba	7.4 ± 1.4	7.0 ± 0.5	5.0 ± 1.0	20 ± 1.3	9.9 ± 1. 1
Cr	1.6 ± 0.8	2.7 ± 0.2	2.3 ± 0.2	7.3 ± 0.4	3.5 ± 0.7
Co	1.7 ± 1.6	1.4 ± 0.2	2.0 ± 0.2	2. 1 ± 0.3	1.8 ± 0.9
Cu	2.6 ± 1.0	5.4 ± 0.3	2.8 ± 0.4	5.0 ± 0.8	3.9 ± 0.9
Ni	2.7 ± 1. 1	1.2 ± 0.2	1.7 ± 0.2	2. 1 ± 0.6	1.9 ± 2. 1
Pb	23.9 ± 1.7	14.2 ± 0.4	11.3 ± 2.3	20.8 ± 1.6	17.6 ± 0.4
V	3. 1 ± 0.9	1.8 ± 0. 1	1.3 ± 0. 1	2. 1 ± 0. 1	2. 1 ± 0.5
As	2.6 ± 0.5	6. 1 ± 0.2	4.3 ± 0.2	4.6 ± 0.5	4.4 ± 0.6
∑Elements	2435 ± 515	1341 ± 222	1436 ± 218	2868 ± 330	2020 ± 325
Ⅱ. Inorganic ions (μg m−3)
K+	0.9 ± 0.2	0.4 ± 0.2	0.4 ± 0.2	1.0 ± 0.4	0.5 ± 0.4
NO3−	9.7 ± 4.8	7.0 ± 4.0	5.0 ± 1.8	9.0 ± 6.3	7.7 ± 1.9
SO42−	11.4 ± 6.0	8.5 ± 2.7	10.2 ± 6.0	23.0 ± 7.8	13.3 ± 5.7
NH4+	9.6 ± 4.6	4.8 ± 1.4	5.4 ± 3. 1	7.0 ± 4. 1	6.7 ± 1.9
∑SNA	30.7 ± 8.9	20.3 ± 6.0	20.5 ± 6.3	38.9 ± 12.8	27.6 ± 7.7
Ⅲ. Gaseous pollutants (μg m−3)
SO2	11.7 ± 6.6	4.9 ± 2.4	9.9 ± 12.5	16.5 ± 12.7	11. 1 ± 11. 1
CO	0.7 ± 0. 1	0.7 ± 0. 1	1 ± 0.5	1 ± 0.6	0.9 ± 0.4
NO2	44.4 ± 12.6	23.3 ± 6. 1	23.6 ± 9. 1	53.8 ± 11.6	37 ± 17.3
O3	61.6 ± 22.7	80 ± 27.6	57 ± 17. 1	38 ± 40.9	57.2 ± 34.8
IV. Meteorological parameters
Temperature (℃)	14.5 ± 2.3	29.4 ± 2.3	21.9 ± 2.8	2.9 ± 4.5	16.4 ± 11.2
Relative humidity (%)	62.2 ± 18.0	75.9 ± 11.3	80.7 ± 14.0	61.6 ± 18.3	70. 1 ± 9.7
Wind speed (m s−1)	4.2 ± 1.8	3.7 ± 1.2	4.6 ± 1.8	3.7 ± 1.8	4.0 ± 3.7
Planet boundary layer height (m)	1175 ± 197	2297 ± 325	1655 ± 328	571 ± 290	1424 ± 731
Solar radiation (W m−2)	6.7 ± 10.0	11.7 ± 14.8	5.6 ± 10. 1	4.6 ± 7.3	7. 1 ± 11.2
PM2.5 (μg m−3)	80.0 ± 28.3	44.3 ± 10.0	48.6 ± 18. 1	89.0 ± 32.2	65.9 ± 31.5

). In general, seasonal changes in PM_2.5 concentrations are influenced by emission strength, transport, chemical transformations, meteorological factors, and deposition processes [63,64,65]. Liaocheng experienced more frequent precipitation in warm seasons (Figure 1), which can effectively remove atmospheric particulate matter. Additionally, favorable atmospheric dispersion conditions (e.g., a higher planetary boundary layer and stronger wind speed) in warmer seasons (Figure 1) were conducive to the dispersal of air pollutants. Compared to warmer seasons, the common use of coal and biofuels for house heating in colder seasons contributed greatly to elevated PM_2.5 concentrations [30,66]. Similarly to the seasonal variation in PM_2.5, CO, SO₂, and NO_x presented higher concentrations in colder seasons than in warmer seasons (Table 3). Conversely, the O₃ concentration was highest in summer, followed by spring, autumn, and winter (Table 3). O₃ is primarily generated from the photochemical reactions involving volatile organic compounds (VOCs) and NO_x and can be used to evaluate the oxidizing ability of aerosols [67]. Thus, it can be concluded that the oxidizing ability of the summertime aerosols was the strongest.

In this work, 18 elements in PM_2.5 samples were determined in each season (Table 3). The mean concentration of total detected elements (TDEs) was 2020 ± 325 ng m⁻³, contributing 3.1% to PM_2.5 mass. Throughout the whole observation period, Al (667 ± 172 ng m⁻³) had the maximum concentration, followed by Ca, Fe, K, S, and Cl (all above 100 ng m⁻³). Elements including Zn, Mg, Mn, Ti, and Pb varied from 10 ng m⁻³ to 75 ng m⁻³, while eight other elements had concentrations lower than 10 ng m⁻³ (Table 3). The first four elements with the highest concentration were Al, Ca, Fe, and K, which contributed 72.2% to TDEs. Among non-crustal elements, Zn, S, and Cl exhibited the highest concentrations, accounting for 20.4% of TDEs, while Co, Ni, and V had the lowest concentrations, collectively accounting for only 0.3% of TDEs.

Similarly to the seasonal patterns of PM_2.5, TDEs exhibited an obvious seasonal change, with levels descending from winter (2868 ± 330 ng m⁻³) to spring (2435 ± 515 ng m⁻³), autumn (1436 ± 218 ng m⁻³), and summer (1341 ± 222 ng m⁻³) (Table 3). This pattern of change was also found in other cities in Shandong Province [4]. The highest concentration of total crustal elements occurred in spring (Table 3), likely due to the dry climate and strong wind conditions, which facilitate dust suspension and dust storm occurrence [68]. The concentrations of K, S, Cl, Zn, Ba, and Cr were notably higher in winter than in other seasons (Figure 2). S primarily originates from coal burning, while Cl and K are largely attributed to biomass combustion [69]. Therefore, the elevated use of coal and biofuels for heating in winter contributed to the higher concentrations of these elements.

To assess the elemental concentration levels in Liaocheng, the regulatory limits from China’s NAAQS for As and Pb [70], the European Air Quality Directives for As, Pb, and Ni [71], and the WHO standard values for V, Pb, and Mn [8] were applied. The concentrations of As (≤6.1 ± 0.2 ng m⁻³) in Liaocheng during four seasons were considerably lower than the average concentration in four cities of Shandong province (0.02 μg m⁻³) [4] and across China (51.0 ng m⁻³) [72]. However, in summer, the As concentration (6.1 ± 0.2 ng m⁻³) approached the limit set by China’s NAAQS (6.0 ng m⁻³) and the EU standard (6.0 ng m⁻³), as well as the WHO guideline (6.6 ng m⁻³), indicating that As was at safe levels except in summer. Additionally, the concentration of Pb (≤23.9 ± 1.7 ng m⁻³) and Ni (≤2.7 ± 1.1 ng m⁻³) were significantly below the limits set by China’s NAAQS (Pb: 500 ng m⁻³) and WHO (Pb: 500 ng m⁻³, Ni: 25 ng m⁻³), as well as the EU directive for Ni (20 ng m⁻³), implying that both elements were also within safe ranges across the four seasons. Moreover, the concentrations of Cr (1.4–7.3 ng m⁻³) were significantly below the mean concentration reported for the other four cities in Shandong province (20.0 ng m⁻³) [4] and across China (85.7 ng m⁻³) [72].

3.2. Enrichment Factors of Elements

EFs are useful for distinguishing between anthropogenic and natural sources and evaluating the extent of anthropogenic influence on elemental concentrations [43]. The EF values of elements such as Mg, Ti, Fe, and Ca were less than 10 during the whole observation period (Figure 3), indicating the dominant contributor of crustal sources (e.g., resuspended soil, rock weathering, and sandstorms) with only slight enrichment [4]. Conversely, the EF values of K, Zn (except in spring, where it reached 105.3), Mn, Ba, Cr, and Ni were distributed in the range of 10’100 across four seasons, suggesting a stronger contribution from anthropogenic sources during most of the observation period. It is worth noting that Co, Cu, Pb, and As were significantly enriched in PM_2.5 with EFs above 100 in four seasons, indicating substantial anthropogenic influence, likely from the metallurgical industry (for Cu) and fossil fuel combustion (for Pb) [73]. Seasonal variations also highlighted different source influences. Fe, K, Mn, and Cr showed the lowest EFs in spring (Figure 3), suggesting a stronger effect from natural crustal sources in spring. In contrast, Ca, Mg, Zn, and Pb had their highest EFs in spring, reflecting an increased influence from anthropogenic emissions.

3.3. Source Appointment

3.3.1. Source Appointment by PMF Analysis

In this work, the PMF model was applied to determine the key sources of elements in PM_2.5 and their respective contributions across different seasons. Four dominating sources were resolved in every season, as displayed in Figure 4.

In spring, the first factor was characterized by the higher percentages of crustal elements such as Ti (70.9%), Mg (63.4%), Al (52.1%), Ca (51.2%), and Fe (46.6%). This strong presence of crustal elements indicates that this factor was regarded as a dust source (Figure 4a). The second factor possessed significant loadings of SO₄²⁻ (49.2%), NO₃⁻ (48.8%), NH₄⁺ (52.7%), and S (41.8%). SO₄²⁻, NO₃⁻, and NH₄⁺ are primarily generated through secondary oxidation reactions such as aqueous-phase reactions and heterogeneous oxidation in the atmosphere [74]. Therefore, this source can be categorized as secondary oxidation. The third factor showed notable contributions from As (62.3%), Zn (59.1%), Pb (55.8%), and Cl (45.2%). As and Cl are well-established markers of coal combustion [4,75]. Following the ban on leaded gasoline in China since 2000, coal burning has become a key emission source of Pb [4]. Abundant As, Zn, and Pb can be released from coal burning in residential areas, as well as from the steel industry and power plants [17,76]. Therefore, the third factor is interpreted as a source related to coal combustion. The fourth factor displayed strong correlations with Mn, Cu, Fe, and Ni. Mn is a significant tracer of iron and steel production, while Cu is an important indicator for copper smelting [17]. Zhang et al. [4] have reported that Cu in Shandong Province is predominantly released from anthropogenic sources (e.g., nonferrous metal industries and the steel industry). Additionally, Ni is proposed as an ideal tracer for oil combustion, often linked to steam boilers and oil-fired power plants [77,78]. Therefore, this fourth factor is identified as industrial emissions.

In summer, the first factor identified by the PMF model represented a secondary oxidation source, characterized by higher percentages of SO₄²⁻ (57.3%), NH₄⁺ (52.4%), and NO₃⁻ (50.1%). (Figure 4b). This indicates that secondary inorganic aerosols are prevalent during this season, likely due to ongoing atmospheric reactions. The second factor presented strong loadings of Fe (55.6%), Mn (51.1%), and Cu (50.2%), suggesting that this factor is indicative of industrial emissions. The third factor was correlated strongly with As (57.3%), Pb (48.2%), Zn (44.7%), and Cl (39.8%). As and Cl are well-established markers of coal combustion [4,75]. Okuda et al. [13] reported that As, Pb, and Zn dominantly originated from coal combustion in the aerosols of Beijing, China. Several studies suggested that Pb was closely linked with coal combustion in urban PM_2.5, China [16,23,79]. Pb in gasoline was forbidden in 2000 all over China. Additionally, the concentration of Pb in fuels should be lower than 0.005 g L⁻¹ in China [16]; thus, Pb has been verified to be largely from coal combustion rather than suspended soil or vehicle emissions [23]. Therefore, this factor could be indicative of coal combustion. The strong relationship with these elements suggests emissions from coal-burning activities, which are notable contributors to air pollution in urban areas. The fourth factor was heavily loaded with V (59.1%), Ba (55.2%), and Ni (48.5%). V in the atmosphere is primarily from fuel oil combustion [80]. High levels of V commonly exist in the lubricating oils of engines as well as fuel and residual oils [81]. V and Ni are particularly abundant in petroleum products [17,82]. Similarly, Liu et al. [16] found that V and Ni were strongly related to the first factor using the PMF model and considered this as a source from fuel oil combustion. Additionally, Ba can be released in significant amounts from tire wear abrasion and brake dust, making it a reliable marker for vehicle emissions [16,83,84]. As a result, the fourth factor was proposed as mixed sources of fuel oil combustion and vehicle emissions.

In autumn, the PMF model identified four main sources, each characterized by specific element contributions. The first factor was abundantly loaded with K⁺ (73.8%), Cl (56.3%), and K (55.6%) (Figure 4e). K⁺ is a typical indicator of biomass burning [42], and a portion of Cl can be released from biomass burning [85]. Thus, this factor can represent biomass burning. The second factor was classified as coal combustion, as evidenced by the high correlations with Zn (72.8%), Pb (58.3%), As (47.6%). These elements are commonly emitted during coal burning processes, reflecting the contributions of this source to PM_2.5 levels in autumn. The third factor showed elevated loadings of Cu (72.8%), Mn (64.2%), Fe (60.9%), and Cr (52.2%), indicating that this factor is primarily linked to industrial emissions. The fourth factor showed higher percentages of V (61.2%), Ni (54.2%), NO₃⁻ (49.5%), and Ba (44.1%), suggesting that this factor represents mixed sources of fuel oil combustion and vehicle emissions. In winter, similarly to autumn, the PMF model also resolved four factors corresponding to biomass burning, industrial emissions, coal combustion, and secondary oxidation (Figure 4f). This consistency across seasons highlights the recurring influence of these sources on PM_2.5 concentrations in Liaocheng throughout the year.

The seasonal differences in the percentage contributions of each detected source are exhibited in Figure 4c–h. In spring, the dust source made the largest contribution (37.2%) to the measured elements (Figure 4c), highlighting the significant impact of dust, particularly during the spring season, when conditions are conducive to dust suspension. In summer, secondary oxidation accounted for the highest percentage (34.8%) in total sources, followed by spring (17.7%) and winter (10.7%) (4c–d, h), primarily because the summertime conditions of higher temperatures and stronger sunshine facilitated the secondary oxidation. The relative abundance of coal combustion was significantly higher in colder seasons, with winter accounting for 53.2% and spring for 25.4%. In contrast, its contributions were lower during the warmer seasons, with summer and autumn each contributing around 16.6%, again suggesting that the enhanced usage of coal for house heating in colder seasons played a crucial role in the deterioration of air quality. The relative contribution of industrial emissions followed a descending order from summer (30.7%) to winter (26.9%), autumn (21.2%), and spring (19.7%). This suggests that there is a need to focus on reducing industrial waste emissions, particularly during the summer months when industrial activities are typically higher. Biomass burning was determined as a key source only in autumn and winter, making up 41.9% and 9.2%, respectively. The substantially high contribution of biomass burning in autumn was largely attributed to agricultural practices, particularly the open combustion of crop waste during the harvest season, especially in October [86,87,88]. Previous studies have suggested that this practice exerts a significant influence on the enhanced PM_2.5 concentration and haze occurrence, especially during the harvest season (from September to October) in the NCP [86,87,88]. In western Shandong Province, intensive crop burning has been observed, leading to severe air pollution. Given that stringent policies on the open burning of agricultural waste (e.g., soybean straw and corn stalk) have been promulgated in Shandong Province since 2013 [88], even more rigorous measures should be enforced during the harvest seasons.

3.3.2. Potential Source Contribution Function (PSCF) Analysis

To investigate the influence of different air masses on the sampling site in Liaocheng, the Hybrid Single-particle Lagrangian Integrated Trajectory model was applied to simulate 48-hour backward trajectories. The Potential Source Contribution Function (PSCF) model was then used to determine the potential source regions of total elements. A comprehensive methodology for this analysis was described by Li et al. [30]. The seasonal variations in the potential source regions for TDEs are exhibited in Figure 5. In spring, the relatively high WPSCF values were mostly located in the vicinity of the observation site, notably in the southwestern and central parts of Shandong Province, including cities such as Taian, Heze, and Jining. Additionally, elevated WPSCF values were found in the eastern part of Henan Province (e.g., Shangqiu and Puyang) and northern Jiangsu Province (e.g., Xuzhou), as well as in parts of Anhui Province (e.g., Suzhou) (Figure 5a). Meanwhile, the high WPSCF values were also observed in Shanxi and Hubei Provinces and Tianjin, suggesting that both long-range transport and local emissions significantly influenced the elemental concentrations in spring (Figure 5a). In summer, the highest WPSCF values of elements were concentrated around the junction of Jiangsu, Anhui, and Shandong Provinces (Figure 5b). Additionally, the median WPSCF values were notably obtained in Rizhao City of Shandong Province and Xiangyang City of Hubei Province. It is worth noting that the air quality in these potential source areas in summer was better than that in Liaocheng (https://www.aqistudy.cn/, accessed on 18 December 2024), indicating that air masses originating from these regions may exert a dilution effect on the local atmosphere in Liaocheng. In contrast, the potential source areas of elements in autumn were predominantly located close to the sampling site (Figure 5c), reflecting the substantial effect of local emissions on the observed elemental concentrations during this season. In winter, the high WPSCF values were mainly distributed in the southwest of Liaocheng, specifically around heavily polluted cities such as Heze in Shandong Province and Kaifeng in Henan Province (Figure 5d). This is consistent with the observation that elemental concentrations during the winter were much higher than those in other seasons, reinforcing the influence of local pollution sources during this season.

3.4. Health Risk Assessment

Although PM_2.5 possesses only a small fraction of certain elements, some are toxic, nonbiodegradable, and carcinogenic [89]. Hence, the carcinogenic and non-carcinogenic risks of selected elements (As, Ni, Co, Pb, Mn, Zn, and V) for different human groups were analyzed (Figure 6). Generally, the carcinogenic risk of the selected elements for adult males was the most significant, followed by adult females and children. However, the HQ value of the selected elements for children was the largest, followed by those for adult males and adult females. This pattern is consistent with previous research by Diao et al. [85], which reported that while the cancer risk for adults was more pronounced than that for children, HQ values tend to be higher in children during haze events. Similarly, Agarwal et al. [90] proposed that although adults face a greater cancer risk, children experience a higher non-cancer risk due to exposure to these pollutants.

TCR exhibited clear seasonal patterns across three demographic groups, with the largest values observed in winter, followed by autumn, summer, and spring (Figure 6a–c). Throughout each season, the cancer risk posed by Co was the most significant for all human groups. Co is a crucial element in coal combustion aerosols [91]; thus, coal combustion should be controlled considering human health. Although coal used in industry and power plants is higher than residential usage, the reduction in residential heating and cooking should be paid greater attention due to the absence of control devices for bulk coal-burning facilities. In addition, stricter controls on coal burning should be further promulgated, and clean energy plans should be continuously implemented around China. Previous studies have reported that the cancer risk posed by As was the most significant during haze periods in Shijiazhuang, China [85]. As has been chosen as a top priority-controlled element in China due to its highest cancer risk across 51 Chinese cities [14]. USEPA [53] suggests that CR values below 10⁻⁶ indicate no significant carcinogenic risk, while values from 10⁻⁶ to 10⁻⁴ indicate an acceptable carcinogenic risk. For children, the CR values of Pb and Ni during the whole observation period ranged from 1.7 × 10⁻⁷ to 2.4 × 10⁻⁶, suggesting both species exerted a low acceptable carcinogenic risk effect on children. However, the CR values for As and Co fell within the range of 1.3 × 10⁻⁶ to 4.3 × 10⁻⁶, indicating that both elements posed a high acceptable carcinogenic risk to children. Likewise, only As and Co possessed average CR values within the range of 10⁻⁶ to 10⁻⁴ for adults in each season, implying that both elements presented a carcinogenic risk to this human group. Moreover, the average TCR values for As and Co in adults (adult males: 1.0 × 10⁻⁵, adult females: 9.6 × 10⁻⁶) were higher than those for children (5.8 × 10⁻⁶), reflecting the more pronounced effect of carcinogenic risk on adults than children. These findings were obtained by Agarwal et al. [90] and Gao et al. [92].

The HI values for each population group presented a distinct seasonal trend, following the pattern of summer > winter > autumn > spring (Figure 6d–f). When the HQ or HI values are below 1, it indicates no noncarcinogenic risk [48]. The HQ values for each metal in every season were consistently lower than the safe threshold (HQ = 1) for all three population groups. Furthermore, the average HI values were 0.30–0.76 for children, 0.14–0.33 for adult males, and 0.13–0.32 for adult females, all of which fell below 1, suggesting that the non-cancer risk in Liaocheng could be considered negligible. This finding was similar to observations in Beijing, China [16].

4. Conclusions

In this work, the pollution characteristics, origins, and health risks of elements in PM_2.5 across four seasons of 2021 were identified to investigate the seasonal variations in elements in the urban atmosphere of Liaocheng. Our analysis revealed the seasonal patterns of PM_2.5 mass, and most elements were characterized by highest concentrations in winter. Additionally, the concentrations of S, Cl, and K were remarkably higher during this season, suggesting that the use of coal and biofuels for house heating exerted a crucial influence on the deterioration of air quality. The EFs for Fe, K, Mn, and Cr were lowest in spring, suggesting the greatest effect of crustal sources on these elements during this season. Conversely, the EFs for Ca, Mg, Zn, and Pb culminated in spring, reflecting the more significant influence of anthropogenic emissions. In spring, the elements were mainly derived from dust sources and coal combustion, contributing 62.6% to the total. In summer, the dominant sources shifted to secondary oxidation and industrial emissions, comprising 65.5% of the total. The greater contribution of coal combustion during autumn and winter was largely due to the enhanced application of coal for house heating. The relative abundance of industrial emissions followed a descending order of summer, winter, autumn, and spring, indicating that attention should be focused on reducing industrial waste, especially in summer. In spite of the implementation of stringent policies on agricultural waste burning in Shandong Province since 2013, biomass burning accounted for the largest percentage (41.9%) in autumn. Therefore, stricter measures should be enforced during the harvest season. The TCR presented the highest values in winter, followed by autumn, summer, and spring. The cancer risk posed by Co was most significant to all demographic groups throughout the observation period. The HI values for each population group presented a downward trend from summer to spring. The HQ values for each metal across all seasons remained below 1 for all three population groups, indicating that the non-cancer risk in Liaocheng could be ignored. These results enhance our understanding of the pollution levels and origins of elements in PM_2.5, providing valuable insights for further model simulations. It should be noted that stricter control policies at regional scales are necessary to mitigate the atmospheric effect of elements and safeguard public health.