Measurement report: Evaluation of sources and mixing state of black carbon aerosol under the background of emission reduction in the North China Plain: implications for radiative effect

Abstract. Accurate understanding of sources and mixing state of black carbon (BC) aerosol is essential for assessing its impacts on air quality and climatic effect. Here, a winter campaign (December 2017–January 2018) was conducted in the North China Plain (NCP) to evaluate the sources, coating composition, and radiative effect of BC under the background of emission reduction since 2013. Results show that liquid fossil fuel source (i.e., traffic emission) and solid fuel source (i.e., biomass and coal burning) contributed 69 % and 31 % to the total BC mass, respectively, using a multiwavelength optical approach combined with the source-based aerosol absorption Ångström exponent values. The air quality model indicates that local emission was the dominant contributor to BC at the measurement site on average, however, emissions in the NCP exerted a critical role for high BC episode. Six classes of BC-containing particles were identified, including (1) BC coated by organic carbon and sulphate (52 % of total BC-containing particles), (2) BC coated by Na and K (24 %), (3) BC coated by K, sulphate, and nitrate (17 %), (4) BC associated with biomass burning (6 %), (5) Pure-BC (1 %), and (6) others (1 %). Different BC sources had distinct impacts on those BC-containing particles. A radiative transfer model estimated that the amount of BC detected can produce an atmospheric forcing of +18.0 W m−2 and a heating rate of 0.5 K day−1. Results presented herein highlight that further reduction of solid fuel combustion-related BC may be a more effective way to mitigate regional warming in the NCP, although larger BC contribution was from liquid fossil fuel source.


distinct among studies (e.g., Kalogridis et al., 2018;Zheng et al., 2019). This may induce a large uncertainty in the accuracy of BC source apportionment. Therefore, a diverse set of AAEs from different source emissions are needed to improve the performance of the aethalometer model.
As an important chemical property of aerosol, BC mixing state describes whether other chemical composition is coated on BC particles (internally mixed) or exists as separate particles (externally mixed). 5 Generally, freshly emitted BC particles (e.g., diesel vehicle emission) exhibit typical external mixing, but over time, they become gradually internally mixed with other non-BC materials (e.g., organics, sulphate, and nitrate) during atmospheric processes (Eriksson et al., 2017). Compared with uncoated BC particles, the coated ones can enhance the absorption efficiency of solar radiation by a factor of 1.2 -2.0, which is strongly associated with the chemical composition of the coatings on BC particles and their thickness 10 (Wang et al., 2014;Fierce et al., 2016;. Although determining BC mixing state through observations is still challenging, emerging advances in online mass spectrometry technique make it possible for obtaining the chemical characteristics of BC coatings. Based on this method, some studies reported direct observations of chemical composition associated with BC particles as well as their evolution features in the atmosphere (e.g., Zhang et al., 2014;Arndt et al., 2017). 15 As a hotspot region for anthropogenic BC emissions, China accounts up to 14% of the global BC radiative forcing (Li et al., 2016). Over the past decade, China has suffered from serious air pollution, especially in the North China Plain (NCP) . To improve air quality, the Chinese State Council has been promulgated a series of regulations to reduce air pollutants, e.g., the most rigorous regulation of the 'Action Plan for the Prevention and Control of Air Pollution' (APPCAP) released on 10 September 2013, 20 which aims to reduce particulate matter by up to 10% by 2017 relative to 2012 levels for all prefecturelevel cities in China. Previous studies have demonstrated the effectiveness of China's clean air actions in reduction of particulate matter (Zhang et al., 2019). The decreased BC and co-emitted pollutants can affect the interactions between BC and secondary aerosols which in turn results in changes in the physicochemical properties of BC aerosol. In the context of emission reduction, assessment of 25 physicochemical characteristics of BC will be useful for improving our understanding of anthropogenic climate impacts in current China. However, studies focused on this aspect are limited at present. Therefore, we conducted intensive measurements during winter in the last year of the APPCAP at a regional site in https://doi.org /10.5194/acp-2020-570 Preprint. Discussion started: 15 July 2020 c Author(s) 2020. CC BY 4.0 License.

Measurement of BC mixing state
The chemical composition of BC coatings was determined by a real-time single particle aerosol mass spectrometer (SPAMS, Hexin Analytical Instrument Co., Ltd., Guangzhou, China). The ambient aerosol is drawn into the evacuated system via a critical orifice (100 µm) with a flowrate of 0.08 L min -1 ; after passing through an aerodynamic lens, the sampled particles are accelerated to certain speeds. Two diode 5 Nd:YAG lasers (MLL-III-532, Changchun, China) operated at 532 nm are used to determine the aerodynamic diameters (0.2 -2.0 µm) of passing particles. After then, a pulsed 266 nm Nd:YAG laser (UL728F11-F115, Quantel, France) is used to ionize those particles. Finally, a dual-polarity time-of-flight mass spectrometer is applied to detect those generated positive and negative fragment ions. A MATLABbased YAADA toolkit (www.yaada.org) was used to search and analyze the imported single-particle mass 10 spectral data. An adaptive resonance theory-based neural network algorithm (ART-2a) was utilized to perform particle clustering, which set a vigilance factor of 0.8, a learning rate of 0.05, and 20 iterations (Li et al., 2019).

Measurement of organic aerosol
The mass concentrations of organic aerosol (OA) in submicron particles were obtained with a aerosol 15 chemical speciation monitor (ACSM, Aerodyne Research Inc., Billerica, MA, USA). Ambient particles were drawn into the ACSM system through a Nafion ® dryer (MD-700-24S-1; Perma Pure, Inc.) to avoid the influence of particle collection efficiency that caused by water condensation in the sampling line. The principle of ACSM has been elaborated elsewhere (Ng et al., 2011). In brief, the sampled particles are focused into a beam through a critical orifice of 100 μm diameter; then those non-refractory particles are 20 vaporized on a hot surface (~600 °C) and ionized with 70 eV electrons; and finally, the mass fragments are detected by a quadrupole mass spectrometer. The primary OA was further resolved into biomass burning OA (BBOA), coal combustion OA (CCOA), and hydrocarbon-like OA (HOA). Detailed information regarding OA source apportionment can be found in supplemental material of Wang et al. (2019b).

Source emission experiments
A custom-made passivated aluminum chamber (~8 m 3 ) was used to characterize the emissions of solid fuels (i.e., biomass and coal). The emitted smokes were diluted by a Model 18 dilution sampler (Baldwin Environmental Inc., Reno, NV, USA) before AE33 aethalometer measurements. The performance evaluation of this chamber is given in Tian et al. (2015). Several residues of alimentary crops (e.g., wheat 5 straw, rice straw, and corn stalk), economic crops (e.g., cotton stalk, sesame stalk, and soybean straw), and firewood were used to represent the biomass burning occurred in the NCP. Each weighted sample was burned on a platform that placed inside the combustion chamber (Wang et al., 2018b). Meanwhile, bituminous coal and honeycomb briquet that commonly used in the NCP were collected in Shanxi and Shaanxi Provinces. A typical stove that extensively used in the NCP was applied for coal test burns inside 10 the combustion chamber (Tian et al., 2019). Furthermore, motor vehicle exhaust emissions were conducted using bench tests. Gasoline and diesel cars at idle speed and several different driving speeds (i.e., 20 and 40 km h -1 ) were tested. 15 An aethalometer model proposed by Sandradewi et al. (2008) was applied to quantify the contributions of liquid fossil fuels (i.e., gasoline and diesel for traffic emission) and solid fuels (i.e., biomass and coal) to BC mass at Xianghe. The measured babs(λ) at the wavelengths of 370 nm (babs(370)) and 880 nm (babs(880)) were used in the model. As demonstrated previously, babs(λ) can be contributed by carbonaceous aerosols and mineral dust. Due to the small mass fraction of mineral dust in PM2.5 during 20 the campaign (11% of PM2.5 mass) and its small mass absorption cross section (0.09 m 2 g -1 at λ = 370 nm and 0.001 m 2 g -1 at λ = 880 nm, Yang et al., 2009), the light absorption caused by mineral dust can be neglected. Previous studies have demonstrated that the babs(880) is mainly contributed by BC aerosol, while babs(370) is associated with BC, primary and secondary brown carbon (pBrC and sBrC, respectively) (Laskin et al., 2015). Finally, the babs(370) and babs(880) can be calculated from the perspective of

Regional chemical dynamical model
The local versus regional contributions to BC mass at Xianghe was quantified by the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem). Here BC used as a tracer was added into the WRF-Chem to improve model's operation efficiency (Zhao et al., 2015). More detailed descriptions regarding the model configurations are shown in Text S1. The performance of WRF-Chem simulation was evaluated with a mathematical parameter of index of agreement (IOA), which describes the relative difference between the simulated and observed values. The IOA can vary from 0 to 1, with the value closer to 1 meaning the better performance of the model simulation. The IOA is calculated as follows (Li et al., 2011): where 9 and 9 are the simulated and observed BC loadings, respectively; !:; and !:; represent the average value of simulated and observed BC loadings, respectively; and denotes the number of simulations. were used as input parameters in OPAC to reconstruct the AOD, SSA, and ASP. The difference between 25 the reconstructed SSA by OPAC and measured SSA by PAX was 1.4%, indicating a reasonable reconstruction result.

Estimations of radiative forcing and heating rate
Further, the atmospheric heating rate ( <= <> , in unit of K d -1 ) induced by BC was calculated using the first law of thermodynamics and hydrostatic equilibrium as follows (Liou, 2002): where ? ( / describes the lapse rate, of which g is the acceleration due to gravity and C C represents the specific heat capacity of air at a constant pressure (1006 J kg -1 K -1 ); ∆F is the atmospheric forcing induced 5 by BC aerosol; and ∆P is representative of the atmospheric pressure difference, which was assumed to be 300 hPa.

Determination of source-specific AAEs
10 Table 1 shows the characteristics of AAEs obtained from source experiments of liquid fossil fuels and solid fuels. The average AAElff was 1.3 ± 0.2, with higher values for gasoline car (1.4 -1.5) than those for diesel engine car (1.1 -1.2). Compared with AAElff, larger average value was found for AAEsf (2.8 ± 1.0) with the largest one obtained from honeycomb briquet emissions (4.0 ± 0.9), followed by firewood burning (2.9 ± 0.2), crop residues emissions (2.4 ± 0.4), and bituminous coal emissions (1.1 ± 0.2). The 15 large variabilities in AAEsf were potentially affected by the various types of solid fuels and their burning conditions; for example, AAEsf shows a weak but significant inverse correlation with combustion efficiency at 95% confidence interval (R 2 = 0.14, p < 0.05, Fig. S4).
The obtained average AAElff (1.3) and AAEsf (2.8) were applied in the aethalometer model to obtain the BC source apportionment. The OA subtypes contributed by liquid fossil fuel (i.e., HOA) and solid fuel 20 sources (i.e., BBOA + CCOA) were used to verify the reliability of model results. In the light of the range of source-based AAEs, a series of AAElff and AAEsf were put into the aethalometer model to obtain the mass concentrations of BClff and BCsf. The correlations were then established for BClff versus HOA and BCsf versus BBOA + CCOA (Fig. 1). The variations in AAElff cannot affect the correlation between BClff and HOA at a fixed AAEsf, but their R 2 increased as AAEsf increased before 3; after the AAEsf larger than 3, the R 2 kept constant regardless of the variability in AAEsf. In contrast, at a fixed AAElff, the R 2 between BCsf and BBOA + CCOA was independent of the AAEsf variation. For the used AAElff (1.3) and AAEsf (2.8), the coefficient of determination of BClff versus HOA (R 2 = 0.60) and BCsf versus BBOA + CCOA (R 2 = 0.66) belonged to the upper limit of all the R 2 values obtained from different ranges of AAElff and AAEsf (Fig. 1). Based on the BC source apportionment result, the estimated ratios of HOA/BClff (1.7) and 5 (BBOA + CCOA)/BCsf (8.4) are comparable to those calculated with emission factors (Cheng et al., 2010;Sun et al., 2018). Therefore, the pair of AAElff of 1.3 and AAEsf of 2.8 is a reasonable selection for this study.

Characteristics of BClff and BCsf
The average BC mass concentration was 3.6 ± 4.0 μg m -3 and varied largely from 0.1 to 24.4 μg m -3 10 during the campaign (Fig. 2a). The estimated BClff comprised 69% (2.5 μg m -3 ) of BC loading, which was over two times larger than the contribution of BCsf (31%, 1.1 μg m -3 ) (Fig. 2b). This indicates that traffic emissions were the dominant contributor to BC mass at Xianghe. As shown in Fig. 2c, the diurnal variation of BClff exhibited two peaks at 08:00 and 19:00, which were coincided with the morning and evening rush-hour traffic. Although BCsf also showed two peaks appeared in the same period with BClff, 15 they were affected by the residential cooking activities in surrounding rural areas, where solid fuel is a commonly used household energy (Liu et al., 2016). Both decreased BClff and BCsf were observed in the afternoon owing to the increases of planetary boundary layer height and wind speed (Fig. 2d, data sources see Text S2). In contrast to a rapid decline in BClff after 19:00, BCsf remained at a high level until midnight.
Meanwhile, the BCsf/BC fraction was also increased towards to the night, indicating the enhanced heating 20 activities with the use of solid fuel on cold nights.  Table S1. Liquid fossil fuel source was an absolute major contributor to BC mass at urban area due to the heavy traffic, but the contribution of solid fuel source enhances at rural area, where wood burning is commonly used as a household energy. Furthermore, we can also find that the dominant 25 BC source in winter NCP has been changed from past solid fuel to current liquid fossil fuel, even though uncertainty may be caused by the limited studies in the NCP. This change is probably attributed to the rigorous regulations promulgated by the Chinese State Council, of which the large-scale project of coalto-gas switching has been considered as an effective way to reduce pollutants. In addition, although highemission motor vehicles are also banned, the total number of vehicles in the NCP region increase largely from 38.7 million in 2013 to 60.3 million in 2017 (NBS, 2018). Therefore, the liquid fossil fuel source has become a more important contributor to BC compared with the solid fuel source.

Contribution of regional transport to BC
The period of 2 -23 January 2018 was arbitrary selected to explore the regional contributions to BC loading at Xianghe using the WRF-Chem model. As shown in Fig. S5, the simulated BC mass concertation correlated significantly with the measured value (R 2 = 0.61, p < 0.01), and the IOA was estimated to be 0.72, indicating that BC formation process was succeed in catching by WRF-Chem. To 10 determine the impacts of local emission and regional transport on BC, we set up six possible BC source regions as part of the modelling exercise (Fig. S6), and their location information is summarized in Table   S2. The percent contributions from local emission and regional transport to BC mass are shown in Fig. 4.
Although the contributions from different regions varied from day to day, the average BC mass contribution from local emission (53%) was slightly higher than that from regional transport (47%), 15 including 20% from Beijing, 5% from Tianjin, 11% from NCP, 9% from northern Hebei, and 2% from other regions.
Although the local emission was the largest contributor to BC mass on average for all the simulated days, its contribution exhibited a negative correlation with the BC loading (Fig. 5a), indicating an increasing importance of regional transport when the high BC episode was occurred. Actually, only BC contributions 20 from Tianjin and NCP increased as the BC loading increased (Fig. 5 c and d), suggesting that the south of Xianghe was an important pollution region to large BC mass at the sampling site. Taking the highest BC loading episode of 12 -13 January as an example to explain the regional transport (Fig. 6). On 11 January before the high BC loading episode, strong north-westerly winds prevailed in the north of Xianghe, and about 44% and 32% of the total BC mass were contributed by local emission and Beijing accounting for 83% of the total BC mass, of which the NCP region accounted for 63% (Fig. 4). On 13 January, the winds switched to south over the NCP region but decreased near the sampling site. The contribution of regional transport reduced to 66% of the total BC mass with 40% from NCP region and 15% from Tianjin (Fig. 4). On 14 January after the high BC episode, the winds turned to the northwest and then the mass concentration of BC decreased gradually.

Chemical composition of BC coatings
The chemical composition of BC coatings was determined with a SPAMS. A total of 454433 particles whose mass spectra had obvious BC fragment ions (e.g., m/z ±12, ±24, ±36, ±48, ±60, and so on) were identified as BC-containing particles. Further, six categories including BC-OCSOx, BC-NaK, BC-KSOxNOx, BC-BB, Pure-BC, and BC-others were classified based on their mass spectral features. The 10 average mass spectral pattern of each class is shown in Fig. 7, and the contribution of each class to the total BC-containing particles is summarized in Table 2.
The BC-OCSOx was characterized by obvious organic carbon (OC) signals in the positive mass spectrum (for example, intense signals of 37 C3H + , 39 C3H3 + , and 50 C4H2 + as well as moderate signals of 27 C2H3 + , 51 C4H3 + , and 63 (CH3)2NH2OH + ) and strong sulphate ( 97 HSO4 -) signal in the negative mass spectrum. This 15 group was the largest contributor, constituting 52% of the total BC-containing particles (Table 2), indicating that BC was mainly coated with OC and sulphate. The presences of 43 C2H3O + (a marker denoting the oxidized organics, Gunsch et al., 2018) and 97 HSO4imply that BC-OCSOx was underwent a certain degree of atmospheric aging processes. As shown in Fig. 8, the BC-OCSOx number fraction increased as BCsf increased. In contrast, it dropped when BClff larger than the value of 75th percentile of 20 BClff. This indicates a greater impact of solid fuel source on BC-OCSOx at a high BC loading environment compared with the liquid fossil fuel source. The diurnal variation in BC-OCSOx number fraction exhibited an upward trend at night after 19:00 (Fig. 9), which was attributed to the intensive domestic heating activities in surrounding rural areas.
The BC-NaK exhibited strong signals of 23 Na + and 39 K + in the positive mass spectrum and less intense 25 signals of 26 CN -, 46 NO2 -, 62 NO3 -, and 97 HSO4in the negative mass spectrum. This group was the second largest contributor accounting for 24% of total BC-containing particles (Table 2). Intense signals of BC fragment ions (m/z 24, 36, 48, 60, and 72) were concentrated on the negative mass spectrum, indicating that BC-NaK particles were mainly freshly emitted. Meanwhile, relatively higher signal was found for nitrate than sulphate in the negative mass spectrum, although their signals were low. This is consistent with the motor vehicle emission, which contains substantial nitrogen oxides (May et al., 2014).
Furthermore, the number fraction of BC-NaK enhanced as BClff increased but kept stable with BCsf ( Fig.   5 8). These results demonstrate that BC-NaK was more likely associated with the fresh traffic emission relative to the solid fuel emission.
The BC-KSOxNOx had a strong signal of 39 K + in the positive mass spectrum and intense signals of 46 NO2 -, 62 NO3 -, and 97 HSO4in the negative mass spectrum. This group comprised 17% of total BC-containing particles ( Table 2). The high signal intensities of nitrate and sulphate indicate that BC-KSOxNOx particles 10 suffered substantial aging processes in the atmosphere. The number fraction of BC-KSOxNOx was the only class that increased in the afternoon, which was consistent with the increase of ozone (O3, Fig. 9) measured with an ultraviolet photometric Model 49i O3 analyzer (Thermo Fisher Scientific, San Jose, CA, USA). This indicates that BC particles were easier to coated with sulphate and nitrate in the more oxidation environment. The strong 39 K + ion signal in the BC-KSOxNOx particles may imply a partial BC-BB contributed only 6% to total BC-containing particles (Table 2), and it was characterized by signals of 39/41 K + in the positive mass spectrum and 26 CN -, 46 NO2 -, and 97 HSO4in the negative mass spectrum.
Several levoglucosan signals of 45 CHO2 -, 59 C2H3O2 -, and 73 C3H5O2were also existed in the negative mass spectrum, suggesting a typical biomass-burning characterization. Intense signals of negative ion spectrum 20 of BC (e.g., m/z -24, -36, and -48) and relatively low signals of nitrate and sulphate indicate that BC-BB particles were less aged in the atmosphere. Due to the intensive heating activities, this class had an increased contribution to total BC-containing particles at night (Fig. 9).
Low ion signals of nitrate and sulphate indicate that Pure-BC was freshly emitted. This group contributed 25 minor (1%, Table 2) to the total BC-containing particles. The BC-others was characterized by some metallic signals (e.g., 40 Ca + , 56 Fe + /CaO + , and 62 FeO + ) and aromatic signatures (e.g., 51 C4H3 + , 63 C5H3 + , 77 C6H5 + , and 91 C7H7 + ) in the positive mass spectrum and strong signal of NO2 -(m/z 46) in the negative https://doi.org/10.5194/acp-2020-570 Preprint. Discussion started: 15 July 2020 c Author(s) 2020. CC BY 4.0 License. mass spectrum. This indicates that this type of BC particles experienced a certain degree of aging processes in the atmosphere and internally mixed with metals and aromatic compounds. This class accounted for only 1% of total BC-containing particles (Table 2).

Implication for radiative effect
The variation in BC forcing effect was highly associated with BC burden in each day. As shown in Fig.   5 10, due to reduction in radiative energy reaching the surface through absorbing incoming sunlight, BC had a cooling effect of -13.6 ± 7.0 W m -2 at the ES, ranging from -1.9 to -27.9 W m -2 . In contrast, BC had a warm effect of +4.4 ± 3.0 W m -2 at the TOA, varying from +0.6 to +20.8 W m -2 , indicating a net energy gain. This was partly attributed to the strong BC absorption that can impede the back scattered radiation reaching the TOA. 10 The difference between BC DRF at the TOA and the ES gave the atmospheric forcing (a net atmospheric absorption) of +18.0 ± 9.6 W m -2 , which corresponded to a heating rate of 0.5 K day -1 . The BC DRF in the atmosphere accounted for 86% of total aerosol DRF (+21.0 W m -2 ), suggesting that BC has significant impact on perturbing the Earth-atmosphere radiative balance. The atmospheric heating in conjunction with the surface reduction in solar flux may aggravate the low-level inversion, leading to slowdown of 15 thermal convection and in turn reducing the process of cloud formation (Chou et al., 2002).
As shown in Fig. 10c and d, the mean BC DRF caused by liquid fossil fuel (solid fuel) source was -7.0 W m -2 (-5.4 W m -2 ) at the ES and +2.1 W m -2 (+1.7 W m -2 ) at the TOA, producing an atmospheric forcing of +9.1 W m -2 (+7.1 W m -2 ). Due to stronger BC DRF, the average heating rate of the atmosphere caused by liquid fossil fuel source (0.3 K day -1 ) was 33% more than solid fuel source (0.2 K day -1 ). Although 20 larger BC DRF was found for liquid fossil fuel source, its atmospheric forcing generated per unit BC mass concentration (3.6 (W m -2 ) (μg m -3 ) -1 ) was 81% smaller compared with the solid fuel source (6.5 (W m -2 ) (μg m -3 ) -1 ). This implies that reduction of solid fuel BC may be a more effective way to mitigate regional warming in the NCP, though BC emission from motor vehicles is also needed to be controlled.

Conclusions
The sources, coating composition, and radiative effect of BC were investigated during winter in the last year of the APPCAP at a regional site in the NCP. Based on the source-specific AAEs (AAElff = 1.3 and AAEsf = 2.8), about 69% of BC was contributed by liquid fossil fuel source while the rest 31% was produced by solid fuel source using an aethalometer model. The BClff and BCsf both showed two peaks 5 at 08:00 and 19:00 but with different causes. Due to rigorous regulations, the dominant BC source in winter NCP maybe changed from past solid fuels to current liquid fossil fuels. The WRF-Chem model shows that local emission (53%) was the largest contributor to BC loading on average, followed by Beijing (20%), Tianjin (5%), NCP (11%), northern Hebei (9%), and other regions (2%).
Based on the mass spectral characteristics, BC coated with OC and sulphate (BC-OCSOx) was the largest 10 contributor (52%) to the measured BC-containing particles. The presences of 43 C2H3O + and 97 HSO4imply that BC-OCSOx was underwent a certain degree of atmospheric aging process. The solid fuel source had a greater impact on BC-OCSOx at a high BC loading environment compared with the liquid fossil fuel source. BC coated with Na and K (BC-NaK) was the second largest contributor to total BCcontaining particles (24%), and this class was more likely associated with the fresh traffic emissions 15 relative to the solid fuel emissions. BC coated with K, sulphate, and nitrate (BC-KSOxNOx) accounted for 17% of the total BC-containing particles. This class was partially influenced by biomass-burning emissions and suffered substantial aging process in the atmosphere. The rest classes of BC-BB, Pure-BC, and BC-other contributed minor to total BC-containing particles (1 -6%).
The SBDART model shows that BC can induce a cooling effect of -13.6 W m -2 at the ES and a warming 20 effect of +4.4 W m -2 at the TOA. The difference between BC DRF at the TOA and ES gave the atmospheric forcing of +18.0 ± 9.6 W m -2 , which can produce a heating rate of 0.5 K day -1 . The atmospheric forcing of BC contributed 86% to the forcing caused by total aerosol in the atmosphere, suggesting that BC had a significant impact on perturbing the Earth-atmosphere radiative balance. From the perspective of BC sources, the atmospheric forcing of BC was higher for liquid fossil fuel source than 25 the solid fuel source, but the atmospheric forcing generated per unit BC mass concentration was larger for solid fuel source. This indicates that reduction of solid fuel BC may be a more effective way to further mitigate the regional warming in the NCP compared with BC that emitted from motor vehicles.
Data availability. The data presented in this study are available at the Zenodo data archive https://doi.org/10.5281/zenodo.3923612.
Supplement. The supplement related to this article is available online.
Author contributions. QW and JC designed the campaign. QW and WR carried out the field measurements.
YZ and WR conducted the source experiments. QW, LL, JZ, and JY analyzed the data. QW drafted the 5 paper. All the authors reviewed and commented on the paper.
Competing interests. The authors declare that they have no conflict of interest.