Modeling Investigation of Brown Carbon Aerosol and Its Light Absorption in China

: Brown carbon (BrC) is a type of organic carbon with light-absorbing abilities, especially in ultraviolet (UV) radiation, which could signiﬁcantly contribute to global warming. Observations have shown high BrC concentrations and absorption in China, suggesting potentially large BrC emissions. The potential contribution of fossil fuel combustion to BrC emission has been ignored in most previous studies. Here, we use GEOS-Chem to simulate BrC distribution and absorption in China, accounting for three major primary BrC sources: residential coal and biofuel combustion, vehicle exhausts, and open biomass burning. Based on the literature and related energy consumption data, we estimate the speciﬁc emission ratio of BrC versus BC, and BrC mass absorption efﬁciency (MAE) for each source. Combined with BC emission, total BrC emission in China is then estimated to be 3.42 Tg yr − 1 in 2018, of which 71% is from residential combustion, 14% is from vehicle exhaust, and 15% is from open biomass burning. Residential combustion is the main source of surface BrC in China, accounting for 60% on average, followed by open biomass burning (23%) and vehicle exhaust emissions (17%). There is a clear seasonality in surface BrC concentrations with the maximum in winter (5.1 µ g m − 3 ), followed by spring (2.8 µ g m − 3 ), autumn (2.3 µ g m − 3 ), and summer (1.3 µ g m − 3 ). BrC AAOD at 365 nm ranges from 0.0017 to 0.060 in China, mainly dominated by residential combustion (73%), followed by open biomass burning (16%), and vehicle exhaust emissions (11%). It is also estimated that BrC accounts for 45–67% (52% on average) of total carbonaceous aerosol AAOD at 365 nm, implying an equal importance of BrC and BC regarding the absorption in UV radiation.


Introduction
Brown carbon (BrC) is a type of organic matter that absorbs light in the visible to ultraviolet (UV) ranges with a strong wavelength dependence and a significant enhancement of light absorption towards the UV regions [1][2][3]. In highly polluted areas, BrC can dominate the light absorption of aerosols, and results in a regional direct radiative forcing of up to 3-4 W m −2 [4]. Accounting for BrC warming effect can even change the regional direct radiative forcing of organic carbon (OC) aerosols from cooling effect (without considering BrC) to warming effect [5,6]. On the global scale, the warming effect of BrC can reach 30-70% of black carbon (BC) and 15-40% of CO 2 [7]. BrC can also affect the formation, the microphysical properties, and the lifetime of clouds, either by acting as cloud condensation nuclei and ice nuclei or through heating the cloud layer [8][9][10][11]. Therefore, BrC plays a key role in regional and global aerosol climate effect.
Observations have shown that BrC surface concentrations in China are generally higher than those observed in European countries [12][13][14]. For example, Wang et al. (2013) reported a BrC column burden of 40-90.5 mg m −2 in Winter, Beijing, much higher than the global average burden of 0.65 mg m −2 [5,15]. In addition to the high pollution level, BrC in China also has a relatively stronger light absorption ability. For instance, the mass absorption efficiency at wavelength of 365 nm (MAE 365 ) of water-soluble brown carbon in Beijing is around 0.7-1.8 m 2 g −1 , which is significantly higher than that in the southeastern United States (0.3-0.7 m 2 g −1 ) [16,17]. Studies have also reported much higher absorption coefficients in Beijing (3.7-10 Mm −1 ) and Sanjiang Plain (22 ± 36 Mm −1 ) in China than those in the United States (0.40-0.88 Mm −1 ), Central and Southern Europe (1.1-4.7 Mm −1 ), and South Korea (2.7 ± 1.4 Mm −1 ) [12,[18][19][20][21]. The high BrC level, combined with high MAE and the consequently high absorption coefficients, indicate a potentially significant climate effect of BrC in China.
There are various sources of BrC in the atmosphere, including incomplete combustion of biomass and fossil fuel, as well as secondary formation [2,5,20,22,23]. Biomass burning is recognized as the most important source of BrC on a global scale [2,4,24]. The potential contribution of fossil fuel combustion to BrC emissions has been pointed out recently. For instance, Xie et al. (2017) [23] and Du et al. (2014) [18] reported that vehicle exhausts contain a large amount of BrC. In addition, Sun et al. (2017) [25] reported about 592 Gg BrC (compared with 482 Gg BC) emitted from residential coal combustion.
Due to the lack of an emission inventory of BrC, most previous model studies either neglect BrC or simply equate BrC with the OC emitted from biomass burning [26,27]. However, recent studies have pointed out that fossil fuel combustion could be a large BrC source, especially in China, due to the huge amount of energy consumption [23,25,28]. Besides this, the optical properties of BrC vary largely with sources [2]. Therefore, it is necessary to consider all potential sources of BrC as well as the corresponding optical properties for each source to improve the understanding of the climate effects due to BrC.
In this study, we use a chemical transport model GEOS-Chem to simulate BrC distribution and its light absorption in China. By considering the emissions from residential coal and biofuel combustion, vehicle exhaust, and open biomass burning combined with the corresponding optical properties for each source, the study aims to (1) investigate the pollution level of BrC in China and (2) improve the estimate of light absorption due to BrC compared with BC in China.

Model Description
We used a chemical transport model GEOS-Chem (version 12.0, http://geos-chem.org, accessed on 2 May 2021) to simulate BrC distribution and its absorption in China. The GEOS-Chem uses assimilated meteorological data GEOS-FP from NASA Global Modeling and Assimilation Office (GMAO) [29]. GEOS-FP data have an original horizontal resolution of 0.25 • × 0.3125 • and 72 vertical layers, with a top layer of around 0.01 hPa. In order to save computational resources, we reduced the number of vertical layers to 47, mainly by merging layers in the stratosphere. The simulation was performed with a nested domain over China with a horizontal resolution of 0.25 • × 0.3125 • . The boundary conditions for the nested simulation were from the global simulation, with a horizontal resolution of The simulation was conducted for the year of 2018 with a full-chemistry simulation of NO x -O x -Hydrocarbon-aerosol. Simulated aerosols include SO 4 2− , NO 3 − , NH 4 + , carbonaceous aerosols (including BC and OC), dust, and sea salt. The simulation of carbonaceous aerosols in GEOS-Chem, as described by Park et al. (2003) [30], accounted for all the atmospheric processes (i.e., emissions, transport, diffusion, aging, secondary formation of OC, and dry and wet deposition). It was assumed that 80% of BC and 50% of primary OC were emitted as hydrophobic, and then became hydrophilic, with an e-folding time of 1.15 days.

Emissions
For the base simulation, the emissions from anthropogenic sources were from the Multi-resolution Emission Inventory for China (MEIC), based on the year of 2017 (http: Atmosphere 2021, 12, 892 3 of 12 //meicmodel.org/?page_id=89, accessed on 2 May 2021). The inventory includes emissions from agriculture, power plants, industry, residential, and transportation sources. The open biomass burning emissions were from the Global Fire Emissions Database (QFED) v2.4r6, with a time resolution of 1 day [31]. Total BC emissions in China are 1.3 Tg yr −1 , of which 49% is from residential combustion, 23% is from vehicle exhaust, and 4.1% is from open biomass burning. The residential combustion mainly refers to coal and biofuel combustion in household stoves. In order to isolate the contribution of residential combustion, vehicle exhaust, and open biomass burning, a series of sensitivity tests were performed in addition to the base simulation, with each test having one of the above emission sources turned off (see Table 1). The difference between the base simulation and each sensitivity test could thus be attributed to the contribution of the specific emission source.

Treatment of Brown Carbon
Most of the previous model studies either neglect BrC or simply use OC from biomass burning as a proxy of BrC. In this study, we consider three major primary sources of BrC: residential combustion (including coal and biofuel), vehicle exhaust, and open biomass burning, which are currently reported as being potentially large BrC sources [23,28,32]. Due to the lack of emission inventories, the emissions of BrC awee estimated based on the emission ratios of BrC versus BC (E BrC/BC ) combined with BC emissions for different sources. Assuming that the atmospheric process of BrC is similar to BC, the atmospheric concentrations of BrC emitted from each source could be estimated as the product of the corresponding E BrC/BC and BC concentrations, as the simulation of BC in GEOS-Chem is linear (i.e., concentrations are proportional to emissions). The calculation of total BrC concentrations can then be described by Equation (1): where C BrC is total mass concentration of BrC, µgC m −3 ; i represents a certain type of emission sources, c i is the BC mass concentration emitted from source i, and E Brc/BC,i is the emission ratio of BrC versus BC for source i.
The corresponding E Brc/BC for each source used in the study are listed in Table 2. For residential combustion, we first decided on E Brc/BC for residential coal combustion and residential biofuel combustion, respectively, and then obtained the mean value weighted by BC emissions from each source. We used E BrC/BC of 5.9 for biofuel combustion based on the study of Sun et al. (2021) [32] and calculated E BrC/BC for coal combustion based on the usage of bituminous coal chunks, bituminous coal briquettes, anthracite coal chunks, and anthracite coal briquettes combined, with the corresponding E BrC/BC for each coal type adopted from Sun et al. (2017) [25]. Finally, based on BC emissions of 547 Gg yr −1 from residential coal combustion and 435 Gg yr −1 from residential biofuel combustion adopted from Wang et al. (2012) [33], we derived an E BrC/BC of 3.9 for residential combustion. E BrC/BC for open biomass burning (9.9) and vehicle exhaust (1.7) were adopted from the study of Tang et al. (2020) [34]. The absorbing aerosol optical depth (AAOD) due to BrC could be described by Equation (2): where m i is the BrC column burden emitted from source i, gC m −2 ; MAE i represents the mass absorption efficiency of BrC emitted from source i, m 2 gC −1 . The BrC MAE for each source at a wavelength of 365 nm is also listed in Table 2. Similar to E Brc/BC , the MAE for residential combustion is a weighted average of those for residential coal combustion and residential biofuel combustion. We used an MAE of 2.3 for biofuel combustion based on the study of Tang [38] also used an MAE of 4.1 m 2 g −1 and 0.886 m 2 g −1 at 350 nm and 550 nm, respectively, for biomass burning BrC in their model simulation. In addition, we used MAE values (at 365 nm) of 10 m 2 g −1 for BC [39] and 0.087 m 2 g −1 for dust [40] for the estimates of AAOD of carbonaceous and total aerosol.

BrC Emissions in China
The total estimated emission of BrC in China in 2018 is 3.42 Tg yr −1 , of which 71% is from residential combustion, 14% is from vehicle exhaust, and 15% is from open biomass burning (Table 2). Emissions from residential combustion are the highest in winter, especially in January, with a maximum of 0.45 Tg month −1 , but start to decrease in spring. In contrast, emissions from open biomass burning peak in spring, with a maximum of 0.11 Tg month −1 in April. There are no obvious seasonal changes in vehicle exhaust emissions. Figure 1 shows the spatial distribution of BrC emissions in China during the year 2018 from residential combustion, vehicle exhaust, and open biomass burning. BrC emissions from residential combustion concentrated in Central, North and Northeast China. BrC emissions from vehicle exhaust are relatively higher in the North China Plain, Yangtze River Delta and Pearl River Delta, with hotspots in megacities. In contrast, the emissions from open biomass burning are scattered and are relatively higher in Northeast Plain.  Compared with other sources, residential combustion is the main source of BrC in China, accounting for 60% of total surface BrC concentrations in China. This result could be explained by the wide usage of biofuel and coal for cooking and heating, especially in rural regions, combined with high BrC emissions due to low combustion efficiency and little emission control. High contributions (>70%) are generally found in Central China. Vehicle exhaust accounts for 17% of surface BrC on average, with relatively higher contributions found in Beijing, Yangtze River Delta, and Pearl River Delta (>40%). Open biomass burning accounts for 23% of surface BrC on average, with maximums on the northeastern and southwestern border of China (40-90%), probably due to biomass burning in neighboring countries. There is also relatively high open biomass burning contribution in Fujian province, reaching up to 50%. This is consistent with previous studies, which show abundant BrC aerosols transported from Southeast Asia (e.g., Indochina) to southern China, especially in spring, due to intensive biomass burning there, associated with southwesterly wind [11,[41][42][43].

Surface Concentrations in China
Surface BrC concentrations also show a clear seasonality, with a maximum in winter  . The high concentrations in winter are mainly due to the increasing residential combustion for heating, which could be confirmed by the largest contribution from residential combustion (82%) in winter (Table 3). In spring, the contribution of residential combustion reaches its minimum (48%), while the contribution of open biomass burning reaches a maximum of 35% compared with the other seasons. The lowest concentrations in summer could be partly attributed to the lowest emissions from residential combustion occurring in that season. Compared with other sources, residential combustion is the main source of BrC in China, accounting for 60% of total surface BrC concentrations in China. This result could be explained by the wide usage of biofuel and coal for cooking and heating, especially in rural regions, combined with high BrC emissions due to low combustion efficiency and little emission control. High contributions (>70%) are generally found in Central China. Vehicle exhaust accounts for 17% of surface BrC on average, with relatively higher contributions found in Beijing, Yangtze River Delta, and Pearl River Delta (>40%). Open biomass burning accounts for 23% of surface BrC on average, with maximums on the northeastern and southwestern border of China (40-90%), probably due to biomass burning in neighboring countries. There is also relatively high open biomass burning contribution in Fujian province, reaching up to 50%. This is consistent with previous studies, which show abundant BrC aerosols transported from Southeast Asia (e.g., Indochina) to southern China, especially in spring, due to intensive biomass burning there, associated with southwesterly wind [11,[41][42][43].
Surface BrC concentrations also show a clear seasonality, with a maximum in winter (5.1 µg m −3 ), followed by spring (2.8 µg m −3 ), autumn (2.3 µg m −3 ), and summer (1.3 µg m −3 ). The high concentrations in winter are mainly due to the increasing residential combustion for heating, which could be confirmed by the largest contribution from residential combustion (82%) in winter (Table 3). In spring, the contribution of residential combustion reaches its minimum (48%), while the contribution of open biomass burning reaches a maximum of 35% compared with the other seasons. The lowest concentrations in summer could be partly attributed to the lowest emissions from residential combustion occurring in that season.

BrC Absorption in China
Simulated AAOD in GEOS-Chem includes the contribution from BrC, BC, and dust. To evaluate the model results for BrC, we also compared model AAOD at 365 nm (with and without the contribution of BrC) with observed AAOD at 440 nm from AERONET sites (level 2.0 daily data [44]). Due to a limited number of valid observations in the year of 2018, only the data from two AERONET sites in Beijing were utilized for the comparison (Figure 3). The comparison shows a better agreement between AERONET and model AAOD at both sites after including the absorption of BrC resulting in the normalized mean bias (NMB) decreasing from −30~−32% to 14~21%. The slight overestimations could be attributed to the difference in the wavelength as the absorption decreases with the wavelength. As a result, it is expected to be higher at 365 nm than at 440 nm. The correlation coefficient r is similar at Beijing site and increases from 0.56 to 0.64 at Beijing-CAMS site. Simulated contribution of BrC to total AAOD at 365 nm at the two sites is about 22%. The comparison implies the importance of considering BrC absorption with regard of the estimates of AAOD in China. coefficient r is similar at Beijing site and increases from 0.56 to 0.64 at Beijing-CAMS site. Simulated contribution of BrC to total AAOD at 365 nm at the two sites is about 22%. The comparison implies the importance of considering BrC absorption with regard of the estimates of AAOD in China.  The contribution of residential combustion to BrC AAOD is the highest (73% on average), followed by vehicle exhausts (11%), and open biomass burning (16%). This is slightly different from the condition of surface BrC concentrations, which could largely be explained by the difference in the MAE among different sources. As listed in Table 2, BrC MAE from residential combustion (2.1 m 2 gC −1 ) and open biomass burning (2.3 m 2 gC −1 ) is much higher than the one from vehicle exhaust (0.74 m 2 gC −1 ). This results in a relatively lower contribution from vehicle exhaust to BrC AAOD in comparison with the condition for BrC concentrations. North China and the Sichuan Basin are the areas with particularly high contributions from residential combustion (56-84%) and vehicle exhaust (~15%). The contributions of open biomass burning in most areas range from 3.0% to 15% but could reach up to 45% in the northeastern and southwestern border of China.
BrC AAOD also shows a clear seasonality with the maximum in spring (0.018), followed by winter (0.015), autumn (0.011), and summer (0.0091). The high value of AAOD in spring is probably associated with open biomass burning, which peaks in spring (33%)  The contribution of residential combustion to BrC AAOD is the highest (73% on average), followed by vehicle exhausts (11%), and open biomass burning (16%). This is slightly different from the condition of surface BrC concentrations, which could largely be explained by the difference in the MAE among different sources. As listed in Table 2, BrC MAE from residential combustion (2.1 m 2 gC −1 ) and open biomass burning (2.3 m 2 gC −1 ) is much higher than the one from vehicle exhaust (0.74 m 2 gC −1 ). This results in a relatively lower contribution from vehicle exhaust to BrC AAOD in comparison with the condition for BrC concentrations. North China and the Sichuan Basin are the areas with particularly high contributions from residential combustion (56-84%) and vehicle exhaust (~15%). The contributions of open biomass burning in most areas range from 3.0% to 15% but could reach up to 45% in the northeastern and southwestern border of China. compared with the other seasons (Table 3). In addition, the BrC MAE for open biomass burning (2.3 m 2 gC −1 ) is the highest among the three sources. In winter, the contribution of open biomass burning is the smallest (4.8%), while the contribution of residential combustion is the largest (83%).   BrC AAOD also shows a clear seasonality with the maximum in spring (0.018), followed by winter (0.015), autumn (0.011), and summer (0.0091). The high value of AAOD in spring is probably associated with open biomass burning, which peaks in spring (33%) compared with the other seasons (Table 3). In addition, the BrC MAE for open biomass burning (2.3 m 2 gC −1 ) is the highest among the three sources. In winter, the contribution of open biomass burning is the smallest (4.8%), while the contribution of residential combustion is the largest (83%). Figure 5 also shows the contribution of BrC to the AAOD of carbonaceous aerosols at 365 nm in China. The contribution of BrC ranges from 45% to 67%, with an average of 52%. In contrast to the spatial distribution of BrC AAOD, the contribution of BrC to total carbonaceous aerosol AAOD is the smallest in North China and Sichuan Basin, where the contribution of vehicle exhaust is relatively high. High BrC contribution is generally found in areas that are more affected by open biomass burning. The results indicate that BrC and BC are of equal importance regarding the absorption at 365 nm, and thus the absorption due to BrC in China should not be ignored.

Conclusions
BrC present in the atmosphere could not only degrade air qu cantly contribute to global warming due to its absorption of solar ra have shown high BrC concentrations and absorption in China, s large BrC emissions in China. However, previous modelling studi or simply equate BrC with the OC emitted from biomass burning. the GEOS-Chem to simulate BrC distribution and its absorption in BrC emissions from residential coal and biofuel combustion, vehi biomass burning.
Based on the literature and related energy consumption data, w 1.7, and 9.9 and MAE of 2.1 m 2 gC −1 , 0.74 m 2 gC −1 , and 2.3 m 2 gC −1 fo combustion, vehicle exhaust, and open biomass burning, respective emissions, we estimate total BrC emissions of 3.42 Tg yr −1 in China is from residential combustion, 14% is from vehicle exhaust, and mass burning. Surface BrC concentrations in China range from 0.05 with higher values found in East and Central China than in West Ch of surface concentrations is residential combustion, accounting fo China, followed by open biomass burning (23%) and vehicle exhaus surface BrC concentrations also show a clear seasonality, with the m µg m −3 ), followed by spring (2.8 µg m −3 ), autumn (2.3 µg m −3 ), and The peak in winter is mainly because of the heating activity in that The comparison with observed AAOD from two AERONET si improvement in the model results when the absorption of BrC is decreasing from −30~−32% to 14~21% and r increasing from 0.56 AAOD in Chian ranges from 0.0017 to 0.060 at 365 nm, with

Conclusions
BrC present in the atmosphere could not only degrade air quality, but also significantly contribute to global warming due to its absorption of solar radiation. Observations have shown high BrC concentrations and absorption in China, suggesting potentially large BrC emissions in China. However, previous modelling studies generally omit BrC or simply equate BrC with the OC emitted from biomass burning. In this study, we use the GEOS-Chem to simulate BrC distribution and its absorption in China, accounting for BrC emissions from residential coal and biofuel combustion, vehicle exhaust, and open biomass burning.
Based on the literature and related energy consumption data, we derive E BrC/BC of 3.9, 1.7, and 9.9 and MAE of 2. The comparison with observed AAOD from two AERONET sites shows a significant improvement in the model results when the absorption of BrC is included, with NMB decreasing from −30~−32% to 14~21% and r increasing from 0.56 to 0.64. Annual BrC AAOD in Chian ranges from 0.0017 to 0.060 at 365 nm, with high values found in Northeast Plain, North China Plain, and the Sichuan Basin (0.040-0.060), mainly due to the contribution of residential combustion. The contribution of BrC to total carbonaceous aerosol AAOD at 365 nm ranges from 45% to 67%, indicating an equal importance of BrC and BC, and, therefore, the necessity of including BrC for accurate estimates of aerosol radiative forcing.