Modeling the impact of chlorine emissions from coal combustion and prescribed waste incineration on tropospheric ozone formation in China

Chlorine radicals can enhance atmospheric oxidation, which potentially increases tropospheric ozone concentration. However, few studies have been done to quantify the impact of chlorine emissions on ozone formation in China due to the lack of a chlorine emission inventory used in air quality models with sufficient resolution. In this study, the Anthropogenic Chlorine Emissions Inventory for China (ACEIC) was developed for the first time, including emissions of hydrogen chloride (HCl) and molecular chlorine (Cl2) from coal combustion and prescribed waste incineration (waste incineration plant). The HCl and Cl2 emissions from coal combustion in China in 2012 were estimated to be 232.9 and 9.4 Gg, respectively, while HCl emission from prescribed waste incineration was estimated to be 2.9 Gg. Spatially the highest emissions of HCl and Cl2 were found in the North China Plain, the Yangtze River Delta, and the Sichuan Basin. Air quality model simulations with the Community Multiscale Air Quality (CMAQ) modeling system were performed for November 2011, and the modeling results derived with and without chlorine emissions were compared. The magnitude of the simulated HCl, Cl2 and ClNO2 agreed reasonably with the observation when anthropogenic chlorine emissions were included in the model. The inclusion of the ACEIC increased the concentration of fine particulate Cl, leading to enhanced heterogeneous reactions between Cl and N2O5, which resulted in the higher production of ClNO2. Photolysis of ClNO2 and Cl2 in the morning and the reaction of HCl with OH in the afternoon produced chlorine radicals which accelerated tropospheric oxidation. When anthropogenic chlorine emissions were included in the model, the monthly mean concentrations of fine particulate Cl, daily maximum 1 h ClNO2, and Cl radicals were estimated to increase by up to about 2.0 μg m−3, 773 pptv, and 1.5× 103 molecule cm−3 in China, respectively. Meanwhile, the monthly mean daily maximum 8 h O3 concentration was found to increase by up to 2.0 ppbv (4.1 %), while the monthly mean NOx concentration decreased by up to 0.5 ppbv (6.1 %). The anthropogenic chlorine emissions potentially increased the 1 h O3 concentration by up to 7.7 ppbv in China. This study highlights the need for the inclusion of anthropogenic chlorine emission in air quality modeling and demonstrated its importance in tropospheric ozone formation.

Abstract.Chlorine radicals can enhance atmospheric oxidation, which potentially increases tropospheric ozone concentration.However, few studies have been done to quantify the impact of chlorine emissions on ozone formation in China due to the lack of a chlorine emission inventory used in air quality models with sufficient resolution.In this study, the Anthropogenic Chlorine Emissions Inventory for China (ACEIC) was developed for the first time, including emissions of hydrogen chloride (HCl) and molecular chlorine (Cl 2 ) from coal combustion and prescribed waste incineration (waste incineration plant).The HCl and Cl 2 emissions from coal combustion in China in 2012 were estimated to be 232.9 and 9.4 Gg, respectively, while HCl emission from prescribed waste incineration was estimated to be 2.9 Gg.Spatially the highest emissions of HCl and Cl 2 were found in the North China Plain, the Yangtze River Delta, and the Sichuan Basin.Air quality model simulations with the Community Multiscale Air Quality (CMAQ) modeling system were performed for November 2011, and the modeling results derived with and without chlorine emissions were compared.The magnitude of the simulated HCl, Cl 2 and ClNO 2 agreed reasonably with the observation when anthropogenic chlorine emissions were included in the model.The inclusion of the ACEIC increased the concentration of fine particulate Cl − , leading to enhanced heterogeneous reactions between Cl − and N 2 O 5 , which resulted in the higher production of ClNO 2 .Photolysis of ClNO 2 and Cl 2 in the morning and the reaction of HCl with OH in the afternoon produced chlorine radicals which accelerated tropospheric oxidation.When anthropogenic chlorine emissions were included in the model, the monthly mean concentrations of fine particulate Cl − , daily maximum 1 h ClNO 2 , and Cl radicals were estimated to increase by up to about 2.0 µg m −3 , 773 pptv, and 1.5 × 10 3 molecule cm −3 in China, respectively.Meanwhile, the monthly mean daily maximum 8 h O 3 concentration was found to increase by up to 2.0 ppbv (4.1 %), while the monthly mean NO x concentration decreased by up to 0.5 ppbv (6.1 %).The anthropogenic chlorine emissions potentially increased the 1 h O 3 concentration by up to 7.7 ppbv in China.This study highlights the need for the inclusion of anthropogenic chlorine emission in air quality modeling and demonstrated its importance in tropospheric ozone formation.

Introduction
Chlorine radicals (Cl) are highly reactive, playing a significant role in the oxidative chemistry of the troposphere (Faxon and Allen, 2013;Young et al., 2014).Similar to hydroxyl radicals (OH), chlorine radicals can oxidize volatile organic compounds (VOCs), which potentially enhance ozone formation.In general, chlorine radicals are more reactive towards most of the VOCs than hydroxyl radicals.The reaction rate constants of chlorine radicals with many alkanes, aromatics, alcohols, and ethers range typically between 1 and Y. Liu et al.: Modeling the impact of chlorine emissions from coal combustion 2 orders of magnitude greater than the corresponding values for hydroxyl radicals (Aschmann and Atkinson, 1995;Nelson et al., 1990;Wang et al., 2005).Hence, the high reaction rates make the chlorine radicals competitive with the OH radicals though the concentration of chlorine radicals is an order of magnitude or more lower than that of hydroxyl radicals (Wingenter et al., 1999).Chlorine radicals can be produced from photodissociation and oxidation of many of the most common chlorinated organic species, but these reaction rates are generally not fast enough to contribute significantly to the concentrations of chlorine radicals.ClNO 2 , Cl 2 , and HCl are dominant primary chlorine radical sources.Riedel et al. (2012) reported that the relative contributions to Cl radicals from ClNO 2 , Cl 2 , and HCl were approximately 45, 10, and 45 %, respectively, in the Los Angeles regions by using a simple box model with local observation.While photolysis of ClNO 2 and Cl 2 occurs in the morning, the reaction of HCl with OH occurs in the afternoon.The reactions are shown as follows: ClNO 2 (g) + hν → Cl (g) + NO 2 (g) (1) Cl 2 (g) + hν → 2Cl (g) (2) In the troposphere, nitryl chloride is formed primarily by the heterogeneous reaction between N 2 O 5 and Cl − (Eqs.4-7) (Bertram and Thornton, 2009;Roberts et al., 2009), while the latter reactant Cl − is the major product of HCl neutralization (Seinfeld and Pandis, 1998;Pio and Harrison, 1987).Therefore, identification of emission sources and quantification of their contributions to ambient HCl levels are of critical importance for the estimation of the abundance of ClNO 2 and/or Cl radicals.
The major sources of tropospheric HCl in the atmosphere include natural sources from sea salt (Keene et al., 1999) and biomass burning (Andreae et al., 1996) and anthropogenic sources from coal combustion and waste incineration (Mc-Culloch et al., 1999).The global annual emission rates of HCl from sea salt and biomass burning were estimated to be 50 Tg Cl yr −1 (Graedel and Keene, 1995;Keene et al., 1999) and 6 Tg Cl yr −1 (Lobert et al., 1999), respectively.Although the emission rates from natural sources are much higher than their anthropogenic counterparts, they are relatively constant and well estimated.The corresponding anthropogenic emission rates from coal combustion and waste incineration were previously estimated to be 4.6 and 2 Tg Cl yr −1 , respectively (McCulloch et al., 1999).Waste incinerations include open waste incineration (the uncontrolled emissions from both residential and dump waste burning) and prescribed waste incineration (the emission from waste incineration plant).The contribution of open waste incineration to HCl was estimated to be 1 Tg yr −1 in China (Wiedinmyer et al., 2014), while that of prescribed waste was unknown, awaiting further investigation.Molecular chlorine (Cl 2 ) is another important precursor of Cl radicals.However, only a few studies on its emissions are available in the literature.Chang et al. (2002) compiled an emission inventory of Cl 2 and HOCl for Houston and estimated an emission of about 10 4 kg per day in total in southeast Texas.Deng et al. (2014) collected flue gas samples from six pulverized coal boiler units of four coal-fired power plants in China and found that about 3.6 % of chlorine in coal could release in the form of gaseous Cl 2 during combustion.Measurements of Cl 2 in urban environment are also sparse, although some were made in marine air and at polar sunrise in the past (Finley and Saltzman, 2008;Lawler et al., 2011;Spicer et al., 2002;Impey et al., 1997).
Once the chlorine emission inventory was constructed, the effects of the chlorine radicals on tropospheric ozone formation can be assessed with air quality models by incorporating the emission inventory.For example, the simulation results of the comprehensive air quality model with extensions (CAMx) found that the emissions of HCl and HOCl could increase 1 h averaged O 3 concentration by 70 ppbv in very localized areas during morning hours (Chang and Allen, 2006).Furthermore, Sarwar and Bhave (2007) estimated the effect of chlorine emission on ozone formation in the eastern United States through model simulations.They found that the monthly mean daily maximum 1 h ozone mixing ratios could be enhanced by up to 3 ppbv in the Houston area when the anthropogenic emissions of Cl 2 and HOCl and the chlorine from sea salt aerosols were considered.
However, the role of the oxidation of hydrocarbons by Cl radicals on O 3 formation is still unclear in China, which is mainly due to the lack of an up-to-date anthropogenic chlorine emission inventory.For example, the most widely used emission inventory, the Multi-resolution Emission Inventory for China (MEIC), which was developed by Tsinghua University (http://www.meicmodel.org),does not include HCl and Cl 2 emissions.In addition, though a reactive chlorine emission inventory (RCEI) in 1990 from coal combustion and waste burning was developed by McCulloch et al. (1999), covering each country all around the world with a resolution of 1 • × 1 • , it could not represent the present situation in China due to the fast industrial and economic development in recent years.Li et al. (2016) applied the RCEI in the WRF-Chem (Weather Research and Forecasting coupled with Chemistry) model to simulate the air quality in the Pearl River Delta of China.Results from sensitivity experiments showed that the simulated particulate Cl − and ClNO 2 concentrations were highly sensitive to the chlorine emissions.There is hence a need to develop an up-to-date anthropogenic chlorine emission inventory in China in order to better model ClNO 2 production and to quantify its effect on atmospheric chemistry and air quality.Development of an anthropogenic chlorine emission inventory can also help policymakers to propose better strategies in air quality management.
In this study, the Anthropogenic Chlorine Emission Inventory for China (ACEIC) was developed for the first time to include the emissions of hydrogen chloride and molecular chlorine from coal combustion and prescribed waste incineration in China.This emission inventory was then applied to the Community Multiscale Air Quality (CMAQ) modeling system to evaluate the effects of chlorine emissions on photochemical O 3 formation through sensitivity analysis.Simulations were performed for November 2011 and the results derived with and without the ACEIC were compared.Section 2 describes the development of chlorine emissions in China.Section 3 presents the model simulation to quantify the impact of these anthropogenic chlorine emissions on atmospheric oxidation and ozone formation.
2 Chlorine emission inventory for China 2.1 Emission from coal combustion

Coal consumption database
Coal consumption data are needed for estimating the chlorine emissions.We selected 2012 as the base year of this emission inventory.The database of coal consumptions was constructed based on the data from the China Energy Statistical Yearbook (CESY, National Bureau of Statistics, 2013), which include 31 provinces, municipalities, and autonomous regions.A zero value of coal consumption in Tibet was assumed in the database due to unavailable data in the yearbook.In addition, coal combustion in Hong Kong and Taiwan were taken from International Energy Agency energy statistics (IEA, 2012) and were included in the database.Hence a total of 33 regions were included in this inventory (Table 1).Similar to the classification method used in the MEIC emission inventory, in the ACEIC we classified the coal consumption from the CESY into four economic sectors according to their characteristics: (1) power plant sector, including electricity plants, heat plants, and combined heat and power (CHP) plants; (2) industrial sector, including iron and steel, nonferrous metals, and other categories covering large-scale combustion processes; (3) residential sector, including personal consumption in both urban and rural regions; (4) other sector, including agriculture, forestry, animal husbandry, fishery, water conservancy, construction, transport, storage, post, wholesale, retail trade, hotel, restaurants, and other consumption.Columns 2-5 in Table 1 list the coal consumption among different categories in different provinces (or regions), along with various chlorine content in coal as discussed in the following section.In 2012, a total of 3.6 million tons of coal were consumed, with Shandong (0.3 million tons) being the highest consumer and Hainan the lowest (about 9000 t, Tibet not included).

Chlorine contents in coal
Chlorine is enriched in coal to some extent and is volatilized during the coal combustion process.The chlorine content in coal is essential for emission estimation and it can vary from region to region.It was reported that the chlorine content in coal in China ranged from 50 to 500 µg g −1 with an average value of ∼ 220 µg g −1 (Tang and Chen, 2002), lower than most other countries.Meanwhile, average chlorine contents between 200 and 300 µg g −1 in China were also reported (Lu, 1996;Zhao et al., 1999).Chen (2010) reported a wide range of chlorine content (39-637 µg g −1 ) with an average chlorine content of 280 µg g −1 , falling within the values mentioned above.In this study, we chose chlorine contents reported from Chen (2010) to estimate chlorine emissions in all regions except Shanghai, Tianjin, Hong Kong, and Taiwan.For these regions where chlorine contents are not listed in the literature, we estimated the chlorine emissions using the average chlorine content (280 µg g −1 ) in China according to Chen (2010).
Some of the coals consumed in China were imported from other countries which might have different chlorine contents.According to the report of the China Energy Statistical Yearbook (CESY, National Bureau of Statistics, 2013), the total amount of coals imported into China was 288 Tg and the total coal consumption in China was 3526 Tg in 2012.Over 91 % of the coals were domestically produced in China.It is hence concluded that the different chlorine content of the imported coals has a minor influence on the estimation of chlorine emission in China.However, it is difficult to evaluate the extent of this influence.Hence, we estimated the chlorine emission from coal combustion in China using the chlorine content of coal from domestic sources and did not take the different chlorine content of the imported coals into account.

Chlorine emission factors
Chlorine emission factors from coal combustion vary with boilers and removal facilities.Table 2 summarizes the chlorine emission factors depending on the combination of boiler types and pollution control technologies in coal combustion (Jiang et al., 2005).This combination can vary significantly from one sector to another.
The net emission factors (EFs) were estimated by the following Eq.( 8): where i represents the province (municipality, autonomous region), j represents the economic sector, k represents the energy allocation type (type of boiler and control device combination), c represents chlorine contents in coal, R is the chlorine release rate, X is the fraction of energy for a sector (energy allocation ratio), η d is the removal efficiency of dust-removal facilities, and η s is the removal efficiency of  sulfate-removal facilities.The chlorine emission factors in power plant, industry, residential, and other categories were calculated based on the parameters given in Tables 1 and 2. They were then applied to estimate the HCl and Cl 2 emissions from coal combustion.

Development of the emission inventory
The HCl and Cl 2 emissions (Es) from coal combustion were estimated as follows: where M represents coal consumption, MM denotes the ratios of the molar mass of chlorine atom to the molecular weight of HCl and Cl 2 (35.5/36.5 for HCl and 1 for Cl 2 ), and ρ is the chlorine proportion of HCl and Cl 2 in emitted flue gas.The flue gas contains chlorine species in the form of particulate Cl − and gaseous HCl and Cl 2 , which were formed through chemical transformation during coal combustion.An average chlorine proportion of about 86.3 and 3.6 % was reported respectively for HCl and Cl 2 in the flue gas samples, which were collected from six pulverized coal boiler units in four coal-fired power plants in China (Deng et al., 2014).We adopted this ρ (86.3 and 3.6 % respectively for HCl and Cl 2 ) when calculating HCl and Cl 2 emissions.Similar procedures were followed to estimate chlorine emissions in Hong Kong and Taiwan.Table 1 lists the calculated HCl and Cl 2 emissions from coal combustion in each region (columns 7-11 for HCl and columns 12-16 for Cl 2 ).
We employed the same resolution (0.25 • × 0.25 • ) as the one used in the MEIC for the meshed grid to develop the ACEIC.The HCl and Cl 2 emissions from each economic sector were spatially allocated into the center of each grid cell.To allocate the chlorine emissions from power plants, a database of the location of each point source was needed.We constructed the database following the procedures below: the chlorine emissions in each grid cell were determined using the emissions of chlorine in the region, multiplied by a ratio of SO 2 emissions from power plants in MIX (an Asian anthropogenic emission inventory) in each grid cell to the total emissions of SO 2 in the region.MIX was developed for the year 2010 to support the Model Intercomparison Study for Asia Phase III (MICS-Asia III) and the Task Force on Hemispheric Transport of Air Pollution (TF HTAP) (Li et al., 2015).MEIC emission was included in MIX.MIX data had five categories: power plant, industry, residential, transport and agriculture.In ACEIC, the locations and relative amount of chlorine emissions from power plants were assumed to be the same as those of SO 2 emissions from the power plants in MIX, though this hypothesis might lead to a small uncertainty.In this way, the spatial distributions of emissions of HCl and Cl 2 from coal combustion of power plants were then determined.Chlorine emissions from other sectors were spatially allocated based on population in 2012.The chlorine emissions in each grid cell were obtained using the chlorine emissions in the region, multiplied by the ratio of population in each grid cell to the total population in the region.The spatial distributions of HCl and Cl 2 emissions from coal combustion of power plant, industry, residential, and others are shown in Figs.S1 and S2 in the Supplement.The chlorine emission of each sector in eastern China was higher than that in western China.
Y. Liu et al.: Modeling the impact of chlorine emissions from coal combustion 2.2 Emission from prescribed waste incineration

Prescribed waste incineration database
Table 1 also lists the waste incineration from garbage disposal incinerators in each province/city from the China Urban-Rural Construction Statistical Yearbook (CURCSY, National Bureau of Statistics, 2012), which was used to estimate chlorine emissions from prescribed waste incineration.Note that the emissions of chlorine were calculated only for the regions with garbage disposal incinerators (a total of 22 regions in this study).The information (location and daily capacity) on the garbage disposal incinerators was obtained from the Information Platform for Municipal Solid Waste Incineration (www.waste-cwin.org).

The HCl emission factor for prescribed waste incineration
Domestic waste contains chlorine in materials such as vegetable matter, paper, plastic, dry cell batteries, and salt (Lightowlers and Cape, 1988).The chlorine content of municipal waste is 0.5 wt % on average.An unabated emission factor of 2.2 g HCl kg −1 for municipal solid waste was reported by Emmel et al. (1989), lower than that for ordinary household waste (3.5 g HCl kg −1 , D. Holland, personal communication, 1991).We adopted the former value (2.2 g HCl kg −1 ) when estimating the HCl emission from prescribed waste incineration.The Cl 2 emission from prescribed waste incineration was not included due to unavailable data from the literature.The net emission factor (EF) for prescribed waste incineration was estimated according to the following Eq.( 10): where EF raw is the unabated emission factor (2.2 g kg −1 ) and η d and η s are the chlorine removal efficiency of dust-removal facilities and sulfate-removal facilities, respectively.We assumed that the control technology of garbage disposal incinerator was similar to the coal combustion of a power plant, and we hence used the average values of η s and η d data for the power plant sector in Table 2 to yield the HCl emission factor.

Development of the emission inventory
The HCl emission (E) for prescribed waste incineration were estimated as follows: where i represents the province (municipality, autonomous region) and M denotes the amount of prescribed waste incineration.The estimated HCl emissions from prescribed waste incineration in each region are listed in Table 1.The HCl emission at each garbage disposal incinerator was obtained using the emission in the region, multiplied by the ratio of the daily capacity of each waste incineration plant to the total daily capacity of all plants in the region (Table 1).The HCl emissions at all the garbage disposal incinerators were then merged into a 0.25 • × 0.25 • grid cell.The results show that high emissions from prescribed waste burning could be seen around the coastal region of eastern China (Fig. S3).

The HCl and Cl 2 emissions
The ACEIC developed in this study included HCl and Cl 2 emissions from coal combustion and HCl emissions from prescribed waste incineration.Table 1 shows the chlorine emissions of all the regions in China, along with Hong Kong and Taiwan.The HCl and Cl 2 emissions from coal combustion in China in 2012 were estimated to be 232.9 and 9.4 Gg, respectively, and HCl emissions from prescribed waste burning were estimated to be 2.9 Gg. Figure 1a and b  general the HCl emission is almost 20 times higher than that of the Cl 2 emission.The North China Plain, the Yangtze River Delta (YRD), and the Sichuan Basin contributed spatially high chlorine emissions.The highest HCl emission was found in Jiangsu, followed by Sichuan and Hebei provinces (Fig. 2).Chlorine emissions were relatively low in South China, including Guangdong, Hunan, Fujian, Jiangxi, and Hainan, probably due to the low chlorine contents in coal used in those regions.The HCl emission from industry contributed to as high as 68 % of the total emissions, followed by others (12 %), residential (10 %), power plant (9 %), and prescribed waste incineration (1 %).Many industrial processes (e.g., iron and steel processing, nonferrous metals processing, cement production) that need coal burning are included in the industrial sector, leading to the highest source of HCl.

Comparison with other chlorine emission
The RCEI developed by McCulloch et al. (1999) was the only emission inventory for chlorine that included China, containing the chlorine emitted globally from coal combustion and waste incineration in 1990.The ACEIC developed in this study made progress in four ways based on the RCEI.
(1) The ACEIC provides a more comprehensive database of coal combustion and prescribed waste incineration in China.
The data in each province/city in China were taken from the CESY and the CURCSY in this study, which were more detailed than those in the RCEI that were from IEA energy statistics and only included the total amount of coal consumption in China.
(2) The ACEIC has a higher resolution (0.25 • × 0.25 • ) than the RCEI (1 • × 1 • ), providing a higher spatial resolution for regional air quality modeling.(3) There are more emission factors.When estimating emission factors, the ACEIC included the removal rates of chlorine from dust-removal facilities and sulfate-removal facilities in China, while the RCEI did not, leading to higher estimated HCl emission in the RCEI.We estimated about 232.9 Gg HCl emission in China in 2012 in ACEIC, only about one-quarter of that estimated from the RCEI (866.7 Gg). ( 4) The ACEIC accounts for Cl 2 emission.The ACEIC includes Cl 2 emission, which is also emitted during coal combustion in China based on the measurement by Deng et al. (2014).
In the following section, the ACEIC was incorporated into the CMAQ model to simulate the air quality in central and eastern China.It was evaluated by comparing the simulated and observed concentrations of chlorine species.In addition, the effect of chlorine emissions on tropospheric ozone formation was quantified to assess its importance in atmospheric chemistry in China.The refined and updated anthropogenic chlorine emission will help to evaluate the impact of chlorine emission on ozone formation in China.
3 Impact of chlorine emissions on tropospheric ozone formation

Model setting
The CMAQ model was developed by the United States Environmental Protection Agency (US EPA) to approach air quality as a whole by including state-of-the-science capabilities for modeling multiple air quality issues, including tropospheric ozone, fine particles, toxics, acid deposition, and visibility degradation (Byun and Schere, 2006).The latest version (5.1) was used in this study.In this study, anthropogenic and biogenic emissions were both included in the simulation.The MIX emission inventory (Li et al., 2017) was used in the simulation.International shipping emission was taken from the Hemispheric Transport Atmospheric Pollution (HTAP) emissions version 2.0 dataset (Janssens-Maenhout et al., 2015).Biogenic emission was calculated from the Model of Emissions of Gas and Aerosols from Nature (MEGAN) (Guenther et al., 2006).Additionally, sea salt emission was calculated during the simulation in the CMAQ model.The methods for estimating sea salt emission and its impact on aerosol chemical formation can be found in Liu et al. (2015).The SAPRC07TIC mechanism (Carter, 2010;Hutzell et al., 2012;Xie et al., 2013;Lin et al., 2013;Pye et al., 2015) was selected as the gas-phase chemical mechanism in the CMAQ model.ISORROPIA (Fountoukis and Nenes, 2007) was used to model the chemistry of inorganic aerosols.Detailed chlorine chemistry (including Eqs.1-7) was considered in the CMAQ model.
The simulation was performed for November 2011.The spin-up time was 10 days (22-31 October) prior to November 2011.During the simulation period, China was controlled by high-pressure systems most of the time, which hindered the transport and diffusion of air pollutants.
Two experiments were set up to evaluate the impact of chlorine emission on tropospheric ozone formation.One ex-periment included the ACEIC in the model (Base experiment), while the other experiment did not (NoCl experiment).The comparison of the Base and NoCl experiments could help to quantify the impacts of anthropogenic chlorine emissions.To include the ACEIC in the CMAQ model, the chlorine emissions from different economic sectors were temporally allocated in different ways.For the coal combustion from the power plant, industrial, and residential sectors, we distributed the total chlorine emissions between each month according to Wu (2009).In addition, the daily distributions of chlorine emissions from the power plant, industrial and residential sectors were allocated the same way as the allocations of the MIX inventory from the corresponding sectors.For the coal combustion from the other sectors, the total chlorine emission was divided equally into each month, each day, and each hour.Since the burning process of garbage disposal incinerators is similar to that of power plants, we assumed the same monthly and daily variation of prescribed waste incineration as that of the power plant sector.Four typical sites in four different regions were selected to analyze the diurnal variations of chlorine species: Beijing (BJ), Shanghai (SH), Guangzhou (GZ), and Chongqing (CQ).The locations of these sites are shown in Fig. 3.

HCl evaluation
Table 3 presents the comparison of modeled HCl concentrations to the observed values in China from available literature.It should be noted that the modeled and observed HCl concentrations were not paired in time and space.The modeled HCl concentrations in the Base and NoCl experiments in Beijing and Guangzhou were underestimated, while those in Shanghai and Hong Kong were overestimated.However, the modeled HCl concentrations from both experiments reasonably matched the observations in a similar magnitude.The difference between the modeled and observed HCl concentration in Beijing reduced when the ACEIC was incorporated into the model, implying the importance of anthropogenic emissions in this region.

Cl 2 evaluation
Reports on Cl 2 measurements in the atmosphere are sparse in the literature.The Cl 2 concentration was measured to be about 2.3 pptv on average in La Jolla (Finley and Saltzman, 2008) and 2.5-20 pptv with a 2-month mean of 3.5 pptv in Irvine, California.We estimated a monthly average concentration of 1-10 pptv (most of urban regions) in China in this study by incorporating the ACEIC into the CMAQ system for air quality modeling (Fig. 6a), which was reasonable compared to the observed values in North America.Cl 2 concentration was very low when anthropogenic chlorine emission was not included in the model (Fig. S8).2006) * Units are µg m −3 .Note that the observed and model values are not paired in time and space.Model predictions are taken from the general geographic areas of the observed data.

ClNO 2 evaluation
The highest ClNO 2 concentrations in China were observed throughout the Northern Hemisphere in the CMAQ simulation (Sarwar et al., 2014).Up to about 2000 pptv ClNO 2 concentration was measured in Hong Kong during the summer of 2012 (Tham et al., 2014) and in Tianjin during the summer of 2014 (Tham et al., 2016).We estimated a monthly average concentration of up to 1178 pptv ClNO 2 in China (Fig. 4g), which was comparable to the observed values.

Impact of HCl emission
Figure 4a shows the spatial distribution of monthly average HCl concentration in the Base experiment.The HCl concentration over the ocean was higher than that over the land, due probably to the largest proportion of HCl emission from the dechlorination of sea salt aerosols (Graedel and Keene, 1995;Keene et al., 1999).The highest concentration of HCl was found in the South China Sea where sea salt emission was also high due to high wind speed and the downwind location (Fig. S4).The impact of chlorine emissions on HCl concentration is shown in Fig. 4b and c.The inclusion of the ACEIC in the model increased the HCl concentration by up to 1.7 µg m −3 in inland China.The chlorine emissions accounted for up to 85.6 % of the HCl concentration in the Sichuan Basin (Fig. 4c).The dechlorination of sea salt aerosols transported to inland area was also considered an important proportion of HCl concentration, especially in South China and the coastal regions in East China (Fig. S5), where the impact of anthropogenic chlorine emission was low.
The spatial distribution of monthly mean concentrations of fine particulate Cl − is shown in Fig. 4d.A higher concentration was found in the North China Plain and the South China Sea.The concentrations of fine particulate Cl − increased by up to 2.0 µg m −3 when anthropogenic chlorine emissions were included in the model (Fig. 4e).The increase of fine particulate Cl − concentration was attributed to the gas-particle partitioning process of HCl and was sensitive to chlorine emissions, especially in the Sichuan Basin, contributing up to an 89 % increase (Fig. 4f).
The HCl concentrations were found to significantly increase in regions such as the Sichuan Basin and the YRD, consistent with the high anthropogenic chlorine emissions shown in Fig. 1a.However, the increase of HCl concentration was negligible in the North China Plain even though there was also high HCl emission, while surprisingly the concentration of particulate Cl − increased more significantly than that in the Sichuan Basin and the YRD.Volatile acidic species (i.e., HCl) can be partitioned into particles by neutralization reactions (Seinfeld and Pandis, 1998).Chlorine partitioning between gas and particle phases ([Cl − ]/([Cl − ] + [HCl])) was calculated and shown in Fig. 5. Higher chlorine partitioning meant that more HCl was transferred into particulate Cl − .The spatial distribution of chlorine partitioning in the Base and NoCl experiment was almost the same, suggesting that the chlorine emissions had little impact on the rate of gas-particle conversion.Higher chlorine partitioning rates were found in the North China Plain than in other regions in inland China, where NH 3 emission was high (Fig. S6), leading to a significant increase of particulate Cl − concentration when the ACEIC was included in the model.Meanwhile, semi-volatile NH 4 Cl is formed via a reversible phase equilibrium with NH 3 and HCl (Pio and Harrison, 1987).When the HCl emission was included in the model, HCl reacts with NH 3 to produce particulate NH + 4 and Cl − , provided that the NH 3 emission was sufficiently high.High NH 3 emission in the North China Plain accelerated the gas-particle transformation from NH 3 to particulate NH + 4 , leading to a decrease of NH 3 concentration and an increase of NH + 4 concentration in PM 2.5 (up to −1.1 and 1.0 µg m −3 , respectively) (Fig. S7).
The spatial distribution of daily maximum 1 h ClNO 2 concentration is shown in Fig. 4g.The ClNO 2 concentration in the North China Plain, the Sichuan Basin, and the coastline along South China were significantly higher than those in other regions.The reservoir species ClNO 2 was formed through the heterogeneous reaction between Cl − and N 2 O 5 on the aerosol surfaces.High Cl − concentrations would accelerate the heterogeneous reaction rates, leading to enhanced ClNO 2 production.The inclusion of the ACEIC in the model increased the monthly daily maximum 1 h ClNO 2 concentration by up to 773 pptv in the whole domain, especially in the North China Plain and Sichuan Basin (Fig. 4h), highlighting the importance of anthropogenic chlorine emissions in ClNO 2 formation (up to 78.4 % of ClNO 2 production was related to anthropogenic chlorine emissions) (Fig. 4i).

Impact of Cl 2 emission
The spatial distribution of the monthly mean Cl 2 concentration is presented in Fig. 6a.As expected, high concentrations were found in regions with high Cl 2 emissions, including the Sichuan Basin, the YRD, and the North China Plain (Fig. 1b).The Cl 2 concentration was very low (< 3.4 × 10 −3 pptv) when anthropogenic chlorine emissions were not included in the model (Fig. S8).The differences between Base and NoCl experiments (Fig. 6b) showed that Cl 2 was almost all from direct emissions; nearly 100 % of Cl 2 was originated from inland anthropogenic chlorine emissions (Fig. 6c).The  results suggested that anthropogenic chlorine emission was a significant source of Cl 2 , which should be included in air quality modeling in order to accurately model regional air quality.

Impact on Cl radicals
Both ClNO 2 (mainly from the heterogeneous reaction between particulate Cl − and N 2 O 5 ) and Cl 2 (mainly from di- peak at about 08:00 and subsequently dropped substantially in the whole morning and early afternoon until 16:00 due to apparently photolysis.The Cl 2 concentration increased gradually after 16:00 and continued to accumulate during nighttime.The Cl radical concentration peaked in the morning due to the photolysis of Cl 2 and ClNO 2 , while it was mainly formed through the reaction of HCl with OH in the afternoon when Cl 2 and ClNO 2 concentrations were very low.Similar diurnal cycles can be found at all sites; however, the impact of anthropogenic chlorine emissions at each site varied.Guangzhou and Shanghai sites could represent the coastal region, where the predominant sources of chlorine were from sea salt emission, while the Chongqing site could represent the inland region, where the predominant sources of chlorine were from coal combustion and waste incineration.The impact of anthropogenic chlorine emissions in Chongqing was higher than those in other regions. The spatial distribution of Cl radical concentration is shown in Fig. 6d.A higher Cl radical concentration was found in the South China Sea, where high HCl concentration was found (Fig. S4a).The concentration of Cl radicals over the land reached up to 8 × 10 3 molecule cm −3 .The Cl concentrations increased by up to 1.5 × 10 3 molecule cm −3 in the whole domain when anthropogenic chlorine emissions were considered in the model (Fig. 6e).The chlorine emission contributed up to 83.3 % of Cl concentration (Fig. 6f).

Impact of chlorine emissions on tropospheric ozone formation
Atmospheric oxidation of VOCs initiated by the Cl radicals plays an important role in tropospheric ozone formation.The oxidation reaction was accelerated as the Cl concentration increased, leading to an increase of ozone and OH radical concentration.The OH radicals can in turn oxidize NO x to produce particulate NO − 3 , resulting in the decrease of NO x concentration.The monthly mean daily maximum 8 h O 3 concentration was high in South China, the Sichuan Basin, and southwestern China during November 2011 and increased by up to 2.0 ppbv (4.1 %) when anthropogenic chlorine emission was included (Fig. 8a-c), especially in central China.It is also shown that the impact of chlorine emission on 1 h O 3 concentration (Fig. S9) was similar to that of 8 h O 3 .The impact of ACEIC was reasonable compared to the result reported in the Houston area (up to 3 ppbv increase of 1 h O 3 concentration) by Sarwar and Bhave (2007).The NO x concentration was observed to be high in the North China Plain (Fig. 8d).It slightly decreased by up to 0.5 ppbv (6.1 %) when the anthropogenic chlorine emissions were included in the model (Fig. 8e-f).In particular, regions such as the North China Plain and the Sichuan Basin were significantly affected by the chlorine emissions.The NO x concentration decreased, corresponding to the increase of O 3 concentration (Fig. 8b and e).This is attributed to the fact that more ozone production leads to the release of OH radicals when chlorine emissions were included in the model, which results in more oxidation of NO x .The maximum impact of chlorine emissions on 1 h O 3 concentration is shown in Fig. 9.The largest increase of 1 h O 3 concentration was found along the Yangtze River, where the chlorine emission potentially increased the 1 h O 3 concentration by up to 7.7 ppbv.

Conclusions
The ACEIC was developed for the first time, which included HCl and Cl 2 from coal combustion and prescribed waste incineration.The HCl and Cl 2 emissions from coal combustion in China in 2012 were estimated to be 232.9 and 9.4 Gg, respectively, while HCl emission from prescribed waste incineration in China was estimated to be 2.9 Gg.The highest emissions of HCl and Cl 2 were found in the North China Plain, the Yangtze River Delta, and the Sichuan Basin.In the ACEIC, HCl emissions from coal combustion of industry contributed 68 % of the total emission, followed by others, residential, power plants, and prescribed waste incineration.
The modeling results with the ACEIC showed that the simulated HCl, Cl 2 , and ClNO 2 agreed reasonably with the observed values.The inclusion of anthropogenic chlorine emissions in the model increased the concentration of fine particulate Cl − , leading to an enhanced heterogeneous reaction of Cl − with N 2 O 5 , which produced ClNO 2 .Reaction of HCl  The Cl radicals reacted with VOCs and potentially enhanced O 3 concentration.The monthly mean concentration of daily maximum 8 h O 3 increased by up to 2.0 ppbv (4.1 %) when the ACEIC was included in the model.The chlorine emission potentially increased the 1 h O 3 concentration by up to 7.7 ppbv in China.As the precursor of O 3 , the monthly mean concentration of NO x decreased by up to 0.5 ppbv (6.1 %).A significant increase of daily maximum 1 h O 3 was found in central China, corresponding to the region with a significant decrease of NO x .
More attention should be paid to the influence of chlorine emissions.The impact of chlorine emissions on ozone formation might vary from season to season.In the future, other typical months will be simulated and analyzed.In addition, emissions of hydrogen chloride and molecular chlorine not only contribute to the increase of tropospheric ozone concentration, but they also enhance the concentrations of particulate NH + 4 .Further studies should focus on the impact of

Table 1 .
Coal consumption and emissions of hydrogen chlorine and molecular chlorine, listed alphabetically according to regions.

Figure 1 .
Figure 1.Spatial distribution of the emissions of hydrogen chloride (a) and molecular chlorine (b) in the ACEIC.
Figure 3. Modeling domain of WRF/CMAQ and the locations of typical sites: Beijing, Shanghai, Guangzhou, and Chongqing.

Figure 4 .
Figure 4. Spatial distributions of modeled HCl (a), fine particulate Cl − (d), and daily maximum 1 h ClNO 2 (g) average concentrations in November 2011 in the Base experiment; the differences (Base minus NoCl) of modeled HCl (b), fine particulate Cl − (e), and daily maximum 1h ClNO 2 (h) average concentrations in November 2011; the percent changes of the NoCl experiment to the Base experiment of modeled HCl (c), fine particulate Cl − (f), and daily maximum 1 h ClNO 2 (i) average concentrations in November 2011.

Figure 6 .
Figure 6.Spatial distributions of modeled Cl 2 (a) and Cl radical (d) average concentrations in November 2011 in the Base experiment; the differences (Base minus NoCl) of modeled Cl 2 (b) and Cl radical (e) average concentrations in November 2011; the percent changes of the NoCl experiment to the Base experiment of modeled Cl 2 (c) and Cl radical (f) average concentrations in November 2011.

Figure 7 .
Figure 7.Diurnal variations of monthly mean concentrations of ClNO 2 (a), Cl 2 (b), and Cl radicals (c) in Beijing, Shanghai, Guangzhou, and Chongqing in the Base and NoCl experiments.

Figure 8 .
Figure 8. Spatial distributions of modeled daily maximum 8 h O 3 (a) and NO x (d) average concentrations in November 2011 in the Base experiment; the differences (Base minus NoCl) of modeled daily maximum 8 h O 3 (b) and NO x (e) average concentrations in November 2011; the percent changes of the Base experiment to the NoCl experiment of modeled daily maximum 8 h O 3 (c) and NO x (f) average concentrations in November 2011.

Figure 9 .
Figure 9. Spatial distribution of the maximum impact of chlorine emissions on 1 h O 3 concentration in November 2011.

Table 2 .
Emission factors of chlorine from coal combustion in China.
The HCl emission in the study region in 2012 from four economic sectors of coal combustion and from prescribed waste incineration.
show the spatial distribution of the total HCl and Cl 2 emissions, respectively, where similar patterns were found, although in

Table 3 .
A comparison of predicted HCl concentrations for the Base and NoCl experiments to observed data from literature * .