1 Introduction

The Fenwei Plain (FWP) and Beijing-Tianjin-Hebei (BTH) region in the north and Yangtze River Delta (YRD) in the south were three key regions for air pollution control in China, among which Fenwei Plain has showed the highest annual mass concentrations of PM2.5 in recent years (Cao and Cui 2021). While many studies have reported the air pollution situation in the more developed areas in BTH and YRD, only a few studies were conducted in the less developed small and medium-sized cities in FWP. The coal-based energy structure, large amounts of heavy industry, and unique basin topography all contributed to the frequent PM2.5 pollution episodes in FWP, especially in the cold seasons (Li et al. 2022; Jia et al. 2023; Yu et al. 2023; Zhang et al. 2023a, b).

Carbonaceous species are a major component of PM2.5, generally accounting for 20–50% of PM2.5 by mass in urban environment (Yu et al. 2019; Zhang et al. 2023a, b). Carbonaceous species, including organic carbon (OC) and elemental carbon (EC), can pose significant adverse impact on local and regional air quality, climate and human health. For example, high levels of toxic organic compounds such as polycyclic aromatic hydrocarbons (PAHs) and phenols are bound on carbonaceous aerosols (Sun et al. 2022; Ai et al. 2023). And EC plays a vital role in regional haze formation and global warming due to its strong absorption capacity of infrared and visible light (Ramanathan and Carmichael 2008; Lu et al. 2021). In addition, brown carbon (BrC), which absorbs solar radiation in the near-ultraviolet and visible spectral range, can also increase the radiative forcing of the atmosphere and contribute to global warming (Yuan et al. 2020; Jiang et al. 2022).

Residential coal combustion and biomass burning are important sources of carbonaceous aerosols in northern China, especially during the heating season (Yan et al. 2017; Wu et al. 2018; Ni et al. 2021). Under the “Wintertime Clean Heating Scheme in Northern China” that was first proposed in 2017, many areas in northern China have gradually upgraded the heating system with a “coal-to-gas” or “coal-to-electricity” transition (Wang et al. 2020). For example, the urban area of Beijing has completely replaced coal with natural gas for central heating in winter (Ai et al. 2023). Consequently, an obvious decreasing trend in OC and EC concentrations was observed in the “2 + 26 cities” in the Beijing-Tianjin-Hebei region and the surrounding areas in northern China (Ji et al. 2019; Dao et al. 2022; Luo et al. 2022). However, the successful implementation of the “coal-to-gas” or “coal-to-electricity” strategies remains a big challenge for the local administration in the less developed cities and rural areas. In FWP, primary emissions from coal and biomass burning were still the major sources of fine particles and carbonaceous aerosols, which showed significantly enhanced contributions during haze episodes in the cold seasons (Zhang et al. 2021; Li et al. 2022; Liu et al. 2023; Yu et al. 2023).

Yuncheng is a medium-sized basin city located in a river valley in the center of FWP. According to the national air quality report, Yuncheng was ranked as one of the most polluted cities in China (Ministry of Ecology and Environment of the People's Republic of China, 2018–2020). Our previous studies have shown that coal combustion, biomass burning and vehicular emissions were the main sources of PAHs and environmentally persistent free radicals (EPFRs) bound on PM2.5 in the heating period (Sun et al. 2022; Ai et al. 2023; Zhang et al. 2023a, b). Highly time-resolved measurements also showed that organics and nitrate were the two major species in PM2.5 in Yuncheng during the cold season and severe haze episodes, which were enhanced by photochemical oxidation, aqueous-phase processing, and local emissions in Yuncheng (Li et al. 2022). The PM2.5 concentration has gradually decreased since 2018 due to the implementation of “Wintertime Clean Heating Scheme in Yuncheng” (The People’s Government of Yuncheng City, 2018–2022). However, the mean PM2.5 concentration in the cold season still exceeded the Grade II National Ambient Air Quality Standard (NAAQS) of 75 μg m−3 (Fig. S1; China National Environmental Monitoring Center, 2015–2022), which could be attributed to the prevalent residential coal combustion and biomass burning in Yuncheng in the cold season.

In this study, day and night PM2.5 samples were collected in Yuncheng in the winter and summer of 2020–2021 to investigate the seasonal variations, sources, and regional transport of carbonaceous species in PM2.5. In particular, the role of domestic heating in the pollution of carbonaceous aerosols was explored by comparison between the before-heating (November 1st to 14th) and heating (November 16th to December 29th) periods. The pollution situation of carbonaceous aerosols in Beijing in winter was also investigated simultaneously, in order to shed light on the effective control strategies of carbonaceous aerosols in the less developed areas in northern China.

2 Materials and methods

2.1 Sampling method

The PM2.5 samples in Yuncheng were collected from November 1st to December 29th, 2020 in winter (including the before-heating period from November 1st to 14th and heating period from November 16th to December 29th) and from July 1st to August 6th, 2021 in summer, with day and night samples from 7:30 to 19:00 and 19:30 to 7:00 the next day, respectively. The winter sampling site was located in Yudu Park (35.06°N, 111.06°E), about 1.5 m above the ground; and the summer sampling site was around 2 km away from the winter sampling site and located on the roof of the Yuncheng Municipal Bureau of Ecology and Environment (35.05°N, 111.05°E), about 40 m above the ground (Fig. S2). The different sampling heights at the two sites would not impair the comparison of samples because the vertical variation of PM2.5 concentration is negligible within the atmospheric mixing layer (Han et al. 2015). Both sites were surrounded by residential buildings and commercial facilities, representing a typical urban environment. The PM2.5 samples in Beijing were collected from November 1st to December 30th, 2020 in winter. The PM2.5 samples in both Yuncheng and Beijing were collected using a middle-volume sampler loaded with a prebaked quartz fiber filter (QFF, Whatman, 90 mm) at a flow rate of 100 L min−1. The sampling details can be found in our previous studies which aimed to investigate the sources and health risks of PAHs and EPFRs in Yuncheng (Sun et al. 2022) and Beijing (Ai et al. 2023). In total, 168 PM2.5 samples were collected in Yuncheng, including 26 samples in the winter before-heating period, 82 samples in the winter heating period, and 60 samples in summer; and 59 PM2.5 samples were collected in Beijing in winter, including 14 samples in the before-heating period and 45 samples in the heating period.

The meteorological conditions (temperature, wind speed, and relative humidity) and concentrations of conventional gaseous pollutants (O3, SO2, CO, and NO2) during the sampling period were obtained from the Yudu Park station (winter, 35.06°N, 111.06°E) and Yuncheng Middle School Station (summer, 35.01°N, 111.00°E) in Yuncheng and Guanyuan monitoring station (39.93°N, 116.36°E) in Beijing.

2.2 Sample analysis

The OC and EC contents of PM2.5 were analyzed using a multi-wavelength carbon analyzer (DRI model 2015, Atmoslytic Inc., USA) following the IMPROVE_A protocol. Thermo-optic reflection (TOR) was used as the optical correction method (Chow et al. 2007; Luo et al. 2021). The samples were heated to 140 °C, 280 °C, 480 °C, and 580 °C in a pure He atmosphere to detect OC1, OC2, OC3, and OC4, respectively. Then the samples were heated to 580 °C, 780 °C, and 840 °C in a 98% He/2% O2 atmosphere to detect EC1, EC2, and EC3, respectively. The OC sub-fractions become less volatile and more refractory from OC1-OC4. And EC1-EC3 are carbons in the form of a single substance. The minimum detection limit (MDL) of the instrument was: total OC: 0.43 μgC cm−2; total EC: 0.12 μgC cm−2; total carbon: 0.49 μgC cm−2.

The contents of secondary organic carbon (SOC) in OC were calculated by the minimum OC/EC ratio method (Turpin and Lim 2001):

$$SOC =OC-EC\times {(OC/EC)}_{pri}$$
(1)

in which (OC/EC)pri was the OC/EC ratio of primary emissions and calculated as the minimum ratio of OC/EC during the sampling period.

2.3 Positive matrix factorization (PMF) analysis

The US-EPA PMF 5.0 model was employed to analyze the sources of carbonaceous species in PM2.5 in Yuncheng. OC1-OC4, EC1-EC3, SOC, SO2 and NO2 were selected as source tracers in the model, and the TC concentration was set as the total variable. The details of the PMF analysis were presented in our previous study (Yu et al. 2021). The data uncertainty (Unc) was calculated by the following equations when the input concentration was higher (Eq. 2) or lower (Eq. 3) than the minimum detection limit (MDL):

$$Unc=\sqrt{{\left(Error\;fraction\times Concentration\right)}^{2}+{\left(0.5\times MDL\right)}^{2}}$$
(2)
$$Unc=5/6 \times MDL$$
(3)

The PMF model was run repeatedly with different numbers of factors, and the optimal solution was selected based on the rationality of factor profiles, the reconstruction of total carbon, and the scaled residuals of the inputs (Yu et al. 2021).

2.4 Potential source contribution function (PSCF)

The potential source regions of OC and EC in Yuncheng were identified by potential source contribution function (PSCF) analysis based on the 72-h backward trajectories starting at a height of 500 m above the ground level (Zhang et al. 2017). The height of 500 m was used to represent the height of the atmospheric boundary layer. The trajectory starting time was set at 03, 07, 11, 15, 19 and 23 UTC respectively. The MeteoInfo software was used to analyze the source trajectory of pollutants (Wang 2019, 2014), and the grid resolution was 0.5° × 0.5°.

PSCF is a method to identify the potential source area of a specific pollutant based on the conditional probability function. The PSCF value can be calculated as follows:

$$PSC{F}_{ij}={M}_{ij}/{N}_{ij}$$
(4)

where \({M}_{ij}\) is the number of trajectories with the pollutant concentration exceeding a specific threshold, and \({N}_{ij}\) is the number of all trajectories passing through the (i, j) grid.

The weighted PSCF value (WPSCF) was further calculated by adopting appropriate weight coefficients (\({W}_{ij}\)) to reduce the uncertainty of the PSCF calculation results (Hsu et al. 2003). The calculation of WPSCF in this study is as follows:

$$WPSC{F}_{ij}=PSC{F}_{ij}\cdot {W}_{ij}$$
(5)
$${W}_{ij}=\left\{\begin{array}{cc}1.0& {3N}_{ave}<{N}_{ij}\\ 0.7& {1.5N}_{ave}<{N}_{ij}\le {3N}_{ave}\\ 0.4& {N}_{ave}<{N}_{ij}\le {1.5N}_{ave}\\ 0.2& {N}_{ij}\le {N}_{ave}\end{array}\right.$$
(6)

3 Results and discussion

3.1 Pollution characteristics of carbonaceous species in PM2.5 in Yuncheng

3.1.1 Seasonal variations of atmospheric pollutants in Yuncheng

The concentrations of PM2.5, carbonaceous species, and gaseous pollutants as well as meteorological conditions in Yuncheng during the winter and summer sampling periods in 2020–2021 are shown in Fig. 1. PM2.5 and O3 were the main air pollutants in winter and summer, respectively. The mean concentrations of PM2.5 were 61.1 ± 20.9 μg m−3, 79.0 ± 37.2 μg m−3 and 25.2 ± 5.9 μg m−3 during the winter before-heating, heating and summer sampling periods, respectively. The PM2.5 concentration was greatly enhanced during the heating period, and was much lower in summer. On the contrary, the mean O3 concentrations in the three sampling periods followed the order of summer (122.3 ± 28.1 μg m−3) > winter before heating (65.0 ± 38.1 μg m−3) > winter heating (39.5 ± 20.2 μg m−3). While the other pollutants including PM2.5, SO2, CO, and NO2 showed the lowest concentrations in summer, O3 appeared to be the dominant air pollutant in Yuncheng in summer.

Fig. 1
figure 1

Temporal variations of meteorological conditions and the concentrations of PM2.5, carbonaceous species, and gaseous pollutants in Yuncheng and Beijing during the winter and summer sampling periods in 2020–2021

Full size image

Carbonaceous species including OC and EC took around 30% of the total mass of PM2.5 in Yuncheng in both seasons. The mean OC and EC concentrations were 13.2 ± 3.8 μg m−3, 20.9 ± 12.6 μg m−3, and 6.5 ± 2.2 μg m−3 for OC and 4.2 ± 1.9 μg m−3, 6.7 ± 4.3 μg m−3, and 1.2 ± 1.2 μg m−3 for EC during the before-heating, heating and summer sampling periods, respectively, corresponding to the PM2.5 mass percentages of 22.9%, 26.5%, and 26.2% for OC and 6.7%, 7.9%, and 4.5% for EC, respectively. Both the mass concentrations of OC and EC and their contributions to PM2.5 were significantly enhanced during the heating period compared to other sampling periods, confirming the impact of heating activities on the pollution of carbonaceous species in PM2.5. Moreover, the SOC/OC ratio was used to indicate the degree of regional secondary pollution (Dan et al. 2004) and was also shown in Fig. 1. The mean SOC/OC ratio in Yuncheng was 71.8% in summer, which was much higher than those during the before-heating (46.1%) and heating (47.2%) periods. The strong solar radiation, high temperature and relative humidity, as well as high O3 concentration in summer were all conducive to SOC generation, through either gaseous oxidation with high oxidant concentrations or atmospheric aqueous processing at high RH (Yu et al. 2021).

3.1.2 Comparison of pollution levels between Yuncheng and other cities

Compared to Yuncheng that still highly relies on coal as the main energy source for domestic heating in cold seasons, Beijing has completely replaced all low-quality bulk coal with natural gas for central heating in winter (Ai et al. 2023). Figure 1 also shows the concentrations of atmospheric pollutants as well as meteorological conditions in Beijing in winter for comparison. As shown in Fig. 1 and Table 1, no obvious differences were observed for the concentrations of atmospheric pollutants before and during the heating periods in Beijing, indicating that the impact of heating activities in Beijing was negligible. Besides, most atmospheric pollutants showed much lower levels in Beijing than in Yucheng during the winter sampling periods, although the population and gross regional product (GRP) of Beijing were around 5 and 22 times respectively of those of Yuncheng in 2020 (Table S1). Such relatively low levels of atmospheric pollutants in Beijing could be attributed to the efficient implementation of Clean Air Action Plan in Beijing (Ai et al. 2023; Dao et al. 2022). As an exception, NO2 was the only pollutant that showed significantly higher concentration in Beijing than in Yuncheng. It has been suggested that the increased natural gas burning after the implementation of the coal-to-gas project in Beijing was responsible for the increased NOx (NO and NO2) concentrations during the heating period (Zhao et al. 2020). However, such increase was not observed for NO2 in Beijing during the heating period compared to the before-heating period, and natural gas burning may not be the main reason for the high levels of NO2 in Beijing. Instead, the high levels of NO2 in Beijing compared to Yuncheng could be due to the higher amount of motor vehicles in Beijing (6.57 million, Table S1) than in Yuncheng (1.29 million, Table S1).

Table 1 Comparison of the mean concentrations of carbonaceous species (μg m−3) in PM2.5 in Yuncheng and other northern and southern cities in China
Full size table

While the mean concentrations of OC (9.5 ± 4.5 μg m−3 before heating, 7.7 ± 3.1 μg m−3 during heating) and EC (2.7 ± 1.7 μg m−3 before heating, 2.2 ± 1.0 μg m−3 during heating) in Beijing in winter were much lower than those in Yuncheng, the mean SOC/OC ratios in both cities in winter were similar, which were 45.3% in Beijing and 46.7% in Yuncheng, respectively. In previous studies, the mean SOC/OC ratio in Beijing in winter was over 50%, and was greatly elevated during the haze episodes (Yu et al. 2019, 2021). The relatively low SOC/OC ratios as well as low O3 concentrations in both cities during the winter sampling periods in this study indicated the insignificant contribution of secondary processes to the compound atmospheric pollution in the two cities in winter.

Table 1 also shows the comparison of the mean concentrations of carbonaceous species in PM2.5 in Yuncheng and other northern and southern cities in China, particularly in winter. Overall, the PM2.5, OC, and EC concentrations in the northern cities in winter with intense heating activities were higher than those in the warmer southern cities. The PM2.5, OC, and EC concentrations in Yuncheng in winter were at the medium level compared with those in other northern cities. It is interesting to note that the pollution levels in Beijing, one of the largest cities in China, was relatively low compared to other northern cities, while another megacity in the south, Chongqing, showed the highest pollution levels of carbonaceous species among the southern cities.

3.2 Correlations of carbonaceous species in PM2.5 with meteorological parameters and gaseous pollutants

To find out the affecting factors and potential sources of carbonaceous species in PM2.5 in Yuncheng, Fig. 2 shows the heat map of Spearman correlation coefficients of carbonaceous species with meteorological parameters and gaseous pollutants in the winter non-heating, winter heating and summer sampling periods. The weather conditions exhibited different impacts on the carbonaceous species in the three different periods. In summer, the carbonaceous species showed positive correlations with temperature but negative correlations with relative humidity. Increasing temperature in summer would increase the emissions of biogenic volatile organic compounds and accelerate the formation of SOC, thus increasing the concentrations of OC and SOC. On the other hand, high RH in summer was typically resulted from the rainfall events, which cleared the air of pollutants (He et al. 2015; Wang et al. 2015; Feng et al. 2021b). Contrary to the case in summer, the carbonaceous species showed negative correlations with temperature but positive correlations with relative humidity in the winter non-heating period. The lower temperature and higher RH in winter were closely associated with lower atmospheric mixing height, which would result in the accumulation of atmospheric pollutants (Chen et al. 2017). Previous studies also showed that aqueous-phase processing at high RH contributed highly to the formation of secondary aerosols in severe haze episodes in winter (Yu et al. 2021; Li et al. 2022). In the winter heating period, wind speed was the most significant affecting factor, reducing the carbonaceous species. Meanwhile, the carbonaceous species was negatively correlated with temperature, which was partially attributed to the increased heating activities at low temperature.

Fig. 2
figure 2

Correlations among carbonaceous species in PM2.5, gaseous pollutants, and meteorological factors for the winter non-heating (a, n = 26), winter heating (b, n = 82) and summer (c, n = 60) sampling periods in Yuncheng (*p ≤ 0.05, **p ≤ 0.01). The red and blue colors represent positive and negative correlation coefficients, respectively, and the smaller area of the ellipse means a greater absolute value of r

Full size image

In regards to the correlations with gaseous pollutants, OC and EC showed significant positive correlations with SO2, CO, and NO2 in the winter heating period. The correlations between the carbonaceous species and SO2, CO, and NO2 in the winter non-heating and summer periods were also positive, but weaker than those in the winter heating period. SO2 mainly came from coal combustion and CO and NO2 were highly emitted from motor vehicles. Therefore, the carbonaceous components of PM2.5 in Yuncheng could be highly affected by coal combustion and vehicle exhaust during the heating period in winter.

3.3 Source analysis of carbonaceous species in PM2.5

3.3.1 Relationships of OC and EC in PM2.5

EC is mainly derived from the combustion of carbon-containing fuels with strong stability in the atmosphere, thus is often used to represent anthropogenic emissions. While the sources of OC are much more complex, high correlations of OC and EC could indicate coincident sources of OC and EC (Ji et al. 2019; Dao, et al. 2022). As shown in Fig. 2, the correlation coefficients between OC and EC followed a decreasing order of winter heating (0.95) > winter non-heating (0.74) > summer (0.61). The high correlation during the winter heating period indicated highly overlapping sources of OC and EC with a high contribution from the anthropogenic combustion activities, and the less significant correlation during the summer period could be resulted from the enhanced contribution of secondary particulate OC.

OC is comprised of primary and secondary organic carbon, which can be expressed as OC = (OC/EC)pri × EC + SOC according to Eq. 1. Although SOC is highly variable with the changes of gaseous precursors and weather conditions, the slope of the linear regression between OC and EC can roughly indicate the OC/EC ratio of primary emissions. Figure 3 shows the linear regressions between OC and EC in PM2.5 in Yuncheng over different sampling periods. The slopes of the linear regressions were 2.73, 1.50, 1.18, respectively in the winter heating, winter non-heating, and summer periods, showing a decreasing trend of (OC/EC)pri in the three periods. According to the literature, the particulate OC/EC values were in the range of 2.4–14.5 for biomass burning (Ram and Sarin 2010), 2.5–10.5 for coal combustion (Chen et al. 2006), and 1.0–4.2 for vehicular emission (Schauer et al. 2002), showing a decreasing trend from solid fuel combustion to vehicular emission. Therefore, the (OC/EC)pri values estimated by the linear regressions between OC and EC suggested that solid fuel combustion and vehicular emission could be the main primary emission source of the carbonaceous species in PM2.5 in the winter heating and summer periods respectively.

Fig. 3
figure 3

Linear regressions between OC and EC in PM2.5 in Yuncheng over different sampling periods

Full size image

3.3.2 Source apportionment in different sampling periods by PMF

Different fractions of OC (OC1-4) and EC (EC1-3) as well as SOC, SO2, and NO2 were employed as the primary and secondary source tracers in PMF to resolve the sources of total carbonaceous species in PM2.5 in Yuncheng. Different factor numbers were examined in the PMF model, and the solution of 5 factors showed optimal performance and was eventually selected. The displacement (DISP) and bootstrap (BS) tests for error estimation were further verified to ensure the accuracy of the PMF model (Table S3). Figure S3 shows the source profiles of total carbonaceous species in PM2.5 in Yuncheng during the whole sampling period. OC1 is the most volatile and the least refractory fraction of OC, while OC4 is the least volatile and the most refractory fraction of OC. Typically, OC1 + OC2 and OC3 + OC4 are classified as volatile organic compounds and refractory organic compounds, respectively (Han et al. 2018). In this study, the concentrations of refractory organic compounds were almost 2 times higher than those of volatile organic compounds in winter, while the concentrations of refractory and volatile organic compounds were nearly equal in summer. Previous studies have shown that OC1 mainly comes from biomass burning, OC2 mainly originates from coal combustion and atmospheric secondary processing, OC3 comes from gasoline vehicle exhaust, OC4 and EC1 come from coal combustion and gasoline vehicle exhaust, and EC2 and EC3 are derived from diesel vehicle exhaust (Chow et al. 2004; Turpin and Huntzicker 1991; Watson 1994; Zhang et al. 2018). Factors 1 and 2 were dominated by EC3 and EC1 respectively, thus were identified as diesel vehicle emission and gasoline vehicle emission respectively. Factor 3 showed high loadings of SO2, NO2, and EC2, thus was defined as industry emission. Factor 4 was dominated by SOC and OC2, representing the secondary formation source. Factor 5 was dominated by OC1, OC4, and EC2, which were mainly derived from biomass burning and coal combustion. Hence, Factor 5 was defined as heating related emission.

The contributions of different sources to total carbonaceous species in PM2.5 during different sampling periods were calculated by PMF and shown in Fig. 4. Apparently, the heating related emission (50.9%) was the dominant source of total carbonaceous species in the winter heating period, while vehicular emissions (49.6%), particularly diesel vehicle emission (29.3%), were the dominant source in summer. Such source apportionment result by PMF was consistent with that obtained by the OC/EC regression analysis. Besides, secondary formation was another main source of carbonaceous species in summer (30.3%), but was less important in the winter non-heating (13.7%) and heating (11.4%) periods.

Fig. 4
figure 4

Source contributions to total carbonaceous species in PM2.5 by PMF during different sampling periods (clean: PM2.5 < 75 μg m-3; polluted: PM2.5 > 75 μg m-3)

Full size image

While coal combustion was the main energy supply in both industrial and heating activities in Yuncheng, the residential coal and biomass burning activities for heating in rural areas would result in more severe pollution due to the lower fuel quality and stove burning efficiency compared to industrial activities, thus posing greater health risks (Zhou et al. 2016). To mitigate the atmospheric pollution associated with the heating activities, an upgrade to central heating with clean energy would be effective, as has been proved by the case in Beijing. Previous studies also showed that CO2 and conventional air pollutants could be reduced by 64% in rural heating sector by substituting coal-burning with wind power in household heating (Ruan et al. 2022), and the control of random coal combustion contributed to approximately 60% of the total PM2.5 reduction in the “2 + 26 cities” in the Beijing-Tianjin-Hebei region and the surrounding areas in China (Wang et al. 2020).

3.3.3 Sources of carbonaceous species during pollution episodes

To raise efficient mitigation measures, the characteristics and sources of carbonaceous species in PM2.5 were also investigated and compared in clean (PM2.5 < 75 μg m−3) and polluted (PM2.5 > 75 μg m−3) periods during the sampling campaign. Figure 5 shows the mean concentrations of carbonaceous species and ratios of SOC/OC in clean and polluted periods in winter and summer. While the OC and EC concentrations were greatly enhanced in the polluted period, the increase of the SOC concentration was less significant (Fig. 5 and Table S1). In fact, both the SOC/OC and OC/EC ratios decreased in the polluted period compared to the clean period in winter, indicating enhanced contributions of anthropogenic primary emissions to OC and EC in PM2.5. As shown in Fig. 4, the contribution of the heating related emission was increased from 27.1% (clean) to 42.1% (polluted) in the winter non-heating period and from 31.1% (clean) to 55.7% (polluted) in the winter heating period, again confirming the significant impact of heating activities on the increase of PM2.5 and its carbonaceous components in winter.

Fig. 5
figure 5

Mean concentrations of carbonaceous species and ratios of SOC/OC in clean (PM2.5 < 75 μg m-3) and polluted (PM2.5 > 75 μg m-3) periods in winter and summer (PM2.5 < 75 μg m-3 for all samples in summer)

Full size image

3.4 Impact of regional transport on local air pollution

Yuncheng is a typical basin city located in FWP. The unique basin topography of Yuncheng tends to trap atmospheric aerosols under stagnant weather conditions. To evaluate the impact of regional transport on local air quality, Fig. 6 shows the potential source regions of OC and EC in PM2.5 in Yuncheng over different sampling periods. Overall, the potential source regions of OC and EC overlapped greatly, with the major EC source regions (WPSCFij > 0.5) more confined in the local and neighboring areas. As shown in Fig. 6, the carbonaceous species in PM2.5 in Yuncheng were significantly affected by transboundary transport in the heating period. While local emissions stood out for the pollution of OC and EC during the heating period, the northwestern areas covering a large area in central Shaanxi, eastern Gansu, Ningxia, and western Inner Mongolia also contributed highly to the carbonaceous aerosols in Yuncheng. Particularly, central Shaanxi appeared to be the key source region of OC and EC during the heating period. Coupling the PMF source analysis results, the high carbonaceous aerosol pollution in Yuncheng during the heating period was therefore attributed to both the local and regional heating activities. The region-wide heating activities in northern China may also affect the air quality in Beijing under adverse weather conditions, even if the clean heating activities have been adopted in Beijing for years (Zhang et al. 2017).

Fig. 6
figure 6

Potential source regions of OC and EC in PM2.5 in the near surface layer (500 m) in Yuncheng by WPSCF over different sampling periods

Full size image

The potential source areas of OC and EC before the heating period were much narrower than the heating period, mainly distributed in the local, southwestern and eastern areas, covering local Yuncheng, southeastern Shaanxi, and western Henan. In summer, air masses transported from the surrounding areas in every direction could contribute to the local OC and EC pollution, with western Henan to the south of Yuncheng the key source region.

4 Conclusions and implications

Yuncheng, a resource-based basin city in Fenwei Plain, was ranked among the most polluted cities in China. This study shows that the mass concentrations of PM2.5 and its carbonaceous components, OC and EC, were increased by 29.3%, 58.3% and 59.5% respectively during the heating period compared to the before-heating period, confirming the impact of heating activities on the pollution of carbonaceous aerosols. Source apportionment results by PMF showed that the heating related emission (50.9%) was the dominant source of total carbonaceous species in the winter heating period, particularly during polluted days. Both local and regional heating activities contributed highly to the pollution of carbonaceous aerosols in Yuncheng, which was reversely related to temperature in winter. For comparison, vehicular emissions (49.6%), particularly diesel vehicle emission (29.3%), and secondary formation (30.3%) were the main sources of total carbonaceous species in summer.

Compared with Yuncheng, the impact of heating activities on air pollution in Beijing was negligible. Most atmospheric pollutants except for NO2 showed much lower levels in Beijing than in Yuncheng, despite of the much higher population and GRP in Beijing. The implementation of the coal-to-gas project in Beijing has proved successful in reducing the air pollutants in winter. An obvious decreasing trend in the OC and EC concentrations was also found in the “2 + 26 cities” in the Beijing-Tianjin-Hebei region and the surrounding areas in China owing to the effective implementation of clean energy policies and air pollution control measures (Ji et al. 2019; Dao et al. 2022; Luo et al. 2022).

Although the “Wintertime Clean Heating Scheme in Northern China” was proposed in 2017, it was still difficult for the local administration of the less developed cities to afford the high cost associated with the transition of heating system (Chen et al. 2019; Zhou et al. 2022). Therefore, the local heating policy should be made on the basis of the economic situations. Besides, various perspectives including energy accessibility, affordability and potential environmental benefits should also be considered for choosing proper substitution fuels for the traditional coal (Chen et al. 2019). Nevertheless, the clean residential heating transition in northern China would significantly reduce the carbon emission, improve the air quality, and reduce the adverse impact of air pollution on human health (Xing et al. 2020), which were all worthwhile.