Trends of multiple air pollutants emissions from residential coal combustion in Beijing and its implication on improving air quality for control measures

Yifeng Xue1，2，Zhen Zhou1，Teng Nie1，Kun Wang2，5，Lei Nie1，Tao Pan1，4，Xiaoqing Wu1，Hezhong Tian2，3，Lianhong Zhong1，Jing Li1，Huanjia Liu2，Shuhan Liu2，Panyang Shao2

1. National Engineering Research Center of Urban Environmental Pollution Control，Beijing Municipal Research Institute of Environmental Protection，Beijing 100037，China；

2. State Key Joint Laboratory of Environmental Simulation & Pollution Control，School of Environment，Beijing Normal University，Beijing 100875，China；

3. Center of Atmospheric Environmental Studies，Beijing Normal University，Beijing 100875，China；

4. School of Environmental Science and Technology，Tianjin University，Tianjin 300072，China；

5. Department of Air Pollution Control，Beijing Municipal Institute of Labour Protection，Beijing 100054，China

Graphical Abstract：

Abstract：Residential coal combustion is considered to be an important source of air pollution in Beijing.However，knowledge regarding the emission characteristics of residential coal combustion and the related impacts on the air quality is very limited.In this study，we have developed an emission inventory for multiple hazardous air pollutants（HAPs）associated with residential coal combustion in Beijing for the period of 2000 to 2012.Furthermore，a widely used regional air quality model，the Community Multi-Scale Air Quality model（CMAQ），is applied to analyze the impact of residential coal combustion on the air quality in Beijing in 2012.The results show that the emissions of primary air pollutants from residential coal combustion have basically remained the same levels during the past decade，however，along with the strict emission control imposed on major industrial sources，the contribution of residential coal combustion emissions to the overall emissions from anthropogenic sources have increased obviously.In particular，the contributions of residential coal combustion to the total air pollutants concentrations of PM10，SO2，NOx，and CO represent approximately 11.6%，27.5%，2.8% and 7.3%，respectively，during the winter heating season.In terms of impact on the spatial variation patterns，the distributions of the pollutants concentrations are similar to the distribution of the associated primary HAPs emissions，which are highly concentrated in the rural-urban fringe zones and rural suburb areas.In addition，emissions of primary pollutants from residential coal combustion are forecasted by using a scenario analysis.Generally，comprehensive measurement must be taken to control residential coal combustion in Beijing.The best way to reduce the associated emissions from residential coal combustion is to use economic incentive to promote the conversion to clean energy sources for residential heating and cooking.In areas with reliable energy supplies，the coal used for residential heating can be replaced with gas，solar energy and electricity.In areas with inadequate clean energy sources，low-sulfur coal should be used instead of the traditional raw coal with high sulfur and ash content，thereby slightly reducing the emissions of PM，SO2，CO and other toxic pollutants.

Key words：residential coal combustion；particulate matter；emission inventory；scenario analysis；CMAQ simulation；temporal and spatial variation

1　Introduction

As the capital of China，Beijing is a typical rapidly growing megacity throughout the world.Heavy local and regional primary air pollutants emissions from various anthropogenic sources，as well as the unfavorable geographical conditions for good diffusion，result in severe air pollution and frequently episodes with high PM2.5 concentration and low visibility in Beijing，which is regarded as one of the most polluted cities in the world （Huang et al.，2016；Lv et al.，2016；Tian et al.，2014；WHO，2015）.According to the Beijing Environmental Statements，the annual average mass concentration of PM2.5 was 85.9μg·m-3 in 2014，nearly 2.5 times of the National Ambient Air Quality Standard（NAAQS）（35μg·m-3 as an annual average），indicating that Beijing has very severe fine particle pollution （Sun et al.，2015；Zhang et al.，2015）.

Residential coal combustion is recognized as a significant source of atmospheric pollution that affects the atmospheric chemistry and climate change.The released hazardous air pollutants（HAPs），including PM，BC，OC，EC，SO2，NOx，CO，CO2，and PCDD/Fs，affect both local and regional air quality and pose a serious threat to human health and the environment （Ge et al.，2004；Shen et al.，2010；Shi et al.，2015；Wornat et al.，2001；Zhi et al.，2008）.However，the knowledge on the magnitude of various air pollutants emissions，their temp-spatial variations，and their effects on urban air quality is still quite limited.Therefore，accurate emission estimates for residential coal combustion are of great importance for predicting atmospheric composition and air quality （Edwards et al.，2004；Scott and Scarrott，2011）.

Several studies on residential coal combustion have been conducted using simple quantification descriptions and generalizations（Chang et al.，2015；Li et al.，2016；Li and Xie，2014；Qiu et al.，2014）.However，the emission inventories developed by previous studies considered a limited number of primary air pollutants and were mostly based on simply calculated and summarized statistical data that were not distinguished by combustion area or type of coal，such as non-smoke coal（bulk or honeycomb briquettes）and smoke coal（raw or honeycomb coal）.With respect to the spatial distribution of residential coal combustion，previous studies generally adopted GDP and population as surrogate indexes to apportion the overall emission inventory into grids with varied resolution，and the associated uncertainty of this distribution was quite high.Consequently，the impact and contribution of residential coal combustion on air quality remains poor understanding and with high uncertainty.

The purpose of this study is to systematically analyze and summarize the residential energy structure of Beijing by adopting the best available emission factors（EFs）to compile the emission inventory of 13 types of HAPs from Beijing residential coal combustion for the period of 2000 to 2012，and investigate the effects on air quality by implementing control measures for residential coal combustion.Furthermore，in combination with scenario analysis，we predict the future emissions from residential coal combustion in Beijing，and propose guidelines for mitigating the pollution associated with residential coal combustion and analyze their implications for control measures or policies.

2　Materials and methods

2.1　Study domain

Beijing，the capital city of the P.R. China，is geographically located at 39°56' N and 116°20'E on the northwestern edge of the North China Plain and is surrounded by the Taihang and Yanshan Mountains to the west，north，and northeast（Gao et al.，2015）.Beijing city covers a total area of 16，410.54km2 and is divided into 16 districts，with approximately 211，480，000 residents（BMBS，2014）.The demand for energy in the industrial sectors and residential communities for heating and cooking is very strong.

2.2　Emission inventory calculation

The magnitude and geographical distribution of multiple air pollutants emissions from residential coal combustion are closely related to the geographical position of the coal-burning area（such as urban or rural），coal stove type，type of coal，and the number of households，which can be generally calculated by using the household area，coal consumption by unit area and specific EF for each species，as shown in the following equation（1）：

　　（1）

where HA is the household area（m2），CC is the coal consumption by per unit area，EF is the emission factor for each species from residential coal-burning areas，and i，j，n are parameters that represent the geographical position of the coal-burning area（such as urban or rural），type of coal，and number of households，respectively.

2.2.1　Activity data

Activity data including the energy consumption，the population of urban and rural was obtained from Beijing Statistic Yearbook.The consumption and type of residential coal was obtained by conducting field investigations organized by the Beijing Environmental Protection Bureau（BJEPB）and cross-checked for coincidence and a comprehensive analysis of annual statistics（BMBS，2014）.A database of residential coal combustion was compiled at the household level.We collected and used detailed information on approximately 1.1 million coal-fired households from 3260 towns and villages.This information includes latitude and longitude，family population，heating area，annual or seasonal consumption of coal and coal type.We compiled and verified the activity data and performed a detailed audit of the accuracy and rationality of the recorded data to confirm the data effectiveness and quality.We also verified the reported coal consumption by residential sector by comparing it with the data from the department of statistics for quality control of the activity data.

Fig.1（a）shows the change of rural population and energy consumed from 2000 to 2012 in Beijing.Due to changes in the division of urban and rural areas，some parts of the original rural areas became urban areas，accompanied slowing economic growth，the rural population declined markedly in 2005，and then it remained stable.By contrast，rural coal consumption has generally increased during the past years with growing coal consumption intensity per capita.The temporary intensified control measures during the year of the Beijing Olympic Games might have caused the large decrease in rural resident coal consumption in 2008.

Subsequently，the rural residential coal consumption has remained at approximately 1.5 million tons，which demonstrates that，for many years，rural coal control measures have failed to achieve their goal and that the coal consumption in rural areas still remains at a quite high level.The electricity consumption in rural areas has grown rapidly，possibly due to the substitution of coal with electricity for residential cooking and heating，and increased use of household electric appliances with the improvement of living standards.A large reduction in rural electricity consumption in 2012 was mainly because of adjustments in the division of urban and rural area，which simultaneously leading to the increasing of urban electricity consumption in the same year.Rural areas in the vicinity of Beijing seldom used liquefied petroleum gas（LPG）or natural gas as household fuels before 2004.Since then，natural gas and LPG consumption have gradually increased over time due to governmental promotion and the growth of residents’ income.However，both the magnitude and the proportions of using these fuels remain low.

Fig.1（b）shows the trend of urban population and energy consumption from 2000 to 2012.Along with the increasing urban population and the change of division of urban and rural in 2012，the consumption of electricity and natural gas in urban areas has gradually increased.The consumption of LPG reached a peak in 2004 and then decreased due to the substitution with natural gas and other energy types.Under the implementation of staged air pollution control actions，the Beijing municipal government has made significant efforts to promote clean energy conversion and other control measures，involving the switch from coal to electricity in apartments and bungalows in urban areas.Coal-to-electricity conversion has resulted in the decrease in directly coal-burning consumption in urban areas.The volume of coal consumption fluctuated from 2004 to 2008 mainly because of the high living costs and the serviceability of electricity heating.After 2008，the local government has performed much more enforcement and supervision to promote coal-to-electricity conversion in the cultural preservation and non-cultural preservation districts in Dongcheng and Xicheng Districts，leading to another drop of coal consumption in urban downtown areas.

Fig.2 shows the coal consumption in residential areas and the ratio of residential coal consumption to total coal consumption.The total coal consumption in Beijing reached a peak in 2005.Since then，it has been gradually decreased，and the total coal consumption in 2012 was 16.7%，which is lower than in 2000.Industrial coal consumption decreased the most，followed by coal use for power plants.In contrast，the annual residential coal consumption fluctuated slightly，and the residential coal consumption in 2012 was only 5.7% lower than that in 2000.Residential coal use only accounted for approximately 12% of total coal use.However，compared with coal-fired industrial boilers and power generation boilers，residential coal is mainly burned in the cool winter season for house heating，and coal stoves have no dedicated devices to reduce the emission of various air pollutants from smoke.These pollutants directly enter the atmosphere at a low altitude and thus present a significant risk to the environment and human health.Therefore，controlling residential coal consumption is of great significance to protect both the environment and human health.

2.2.2　Emission factors

The EFs of PM depend on the characteristics of unabated emissions from the overall combustion process.Here，we adopt recently collected field measurements of PM2.5，BC，EC，OC and TOC for Chinese residential coal stoves（Chen et al.，2015）.The ratio of PM2.5/PM10 was determined according to the report of Ge et al.（Ge et al.，2004）.The SO2 data were based on the

sulfur contents and retention fractions of coal.The Beijing municipal government issued a local standard for low-sulfur coal and its products（DB 11/097—2004）in 2004 and revised this standard in 2013 and 2014（DB 11/097—2013，DB 11/097—2014）.As a result，the average sulfur content in coal for residential uses declined from about 0.6% in 2007 to 0.4% in 2013.In this study，we use the calculation equation of EF for SO2，which considers the variation in sulfur content of the coal used in coal stove and the current emission situation.The EFs of other gaseous pollutants，such as NOx，VOCs，PAHs，PCDD/Fs，B[a]P，CO and HCl，for coal stoves were referenced from（Paradiž et al.，2008）.The selected average EFs for various air pollutants are summarized in Table 1.Notably，the EFs of SO2 from 2000 to 2012 change with the annual averaged S content in coal.Whereas，considering that there is no significant variation in the emission control of PM，NOx，CO，VOCs and other pollutants from 2000 to 2012 and the lack of newly local field test results for determining yearly-varying EFs，we choose to use the constant average EFs to represent the overall emission situation during this period.This inevitably introduces some uncertainties into our emission inventories that are discussed in the uncertainty analysis section.

①Sis the sulfur content of the combusted coal（%）.

2.3　Simulation

We applied the Models-3/Community Multi-scale Air Quality（CMAQ version 5.0.2）modeling system to conduct the air quality modeling（Djalalova et al.，2015；Foley et al.，2015；Gan et al.，2015）.The first domain with a 36km resolution covers the entire East Asia.The second domain with a 12km resolution covers North China，including Beijing，Tianjin，Hebei，and parts of Shanxi，Shandong，Liaoning，and Inner Mongolia.The third domain with a 4km resolution focuses on Beijing and includes Tianjin and parts of Hebei as well.The vertical resolution is 13 layers extending from the surface to the tropopause，and the first layer is approximately 18m thick.

The meteorological fields are provided by Weather Research and Forecasting（WRF version 3.6）model.The National Center for Environmental Prediction（NCEP）FNL Global Tropospheric Analysis Data is used to generate the first guess field at 6 hours intervals.The NCEP ADP Global Air Observational Weather Data is used in the objective analysis to improve the first guess field.The main physical parameterizations selected in the WRF model are WSM 3-class simple ice microphysics scheme，RRTM long wave radiation scheme，MM5 Monin-Obukhov surface-layer scheme，thermal diffusion land-surface scheme，MRF planetary boundary layer scheme and Grell 3D ensemble cumulus scheme （Hogrefe et al.，2015；Syrakov et al.，2016；Wang et al.，2015a）.

The Sparse Matrix Operator Kernel Emissions（SMOKE version 3.5.1）model is applied to prepare the emission input for CMAQ （Wang et al.，2015b；Zhang et al.，2016）.The first domain uses MEIC anthropogenic emission inventory with a 0.25°×0.25° resolution（http：//www.meicmodel.org/），and the inner two domains use a local inventory developed by Beijing Municipal Research Institute of Environmental Protection（BMRIEP）which is based on detailed environmental statistics and emission factor method containing seven kinds of primary air pollutants（i.e.，CO，NOX，SO2，PM10，PM2.5，VOCs，NH3）.The results of this local inventory can be found in the report released by the BMRIEP（BMRIEP，2013）.The biogenic emissions of VOCs from vegetation and NO from soil are calculated using the Biogenic Emissions Inventory System（BEIS version 3.14），based on global 30-arc second MODIS land-use data（Fu and Liao，2012；Li and Xie，2014）.

In the model simulation，we replaced the emissions from residential coal combustion with our own results.The simulation periods were January，April，July，and October 2012，which representing winter，spring，summer，and autumn，respectively.Each period started five days in advance to minimize the influence of initial conditions.We used the brute-force method to calculate the contribution.Two simulations were performed.Simulation 1（SIM1）takes the emission of coal-fired stoves into consideration，while Simulation 2（SIM2）does not.The difference between SIM1 and SIM2 represents the contribution of emission from coal-fired stoves to air quality.

In order to evaluate the model performance three statistics variables were calculated to determine the model performance：normalized mean bias（NMB），normalized mean error（NME）and correlation coefficient（COR），defined as.

Where and represent simulated and observed concentrations，respectively.

2.4　Scenario analysis

Pollution control situations are forecasted with scenario analysis for the following three：one Business-As-Usual（BAU）scenario and two Improved Pollution Control（IPC-1，IPC-2）scenarios（Xue et al.，2016）.We assumed that the energy consumption by residential sector would remain stable until 2030 in association with a stable population and GDP of Beijing.However，the consumption of different energy types including coal，natural gas，electricity etc.will vary based on energy and environmental policies.In the BAU scenario，we assume that residential coal combustion will continue following the current emissions model，environmental supervision and enforcement do not improve，and coal consumption by rural residents for heating and cooking increases.In the IPC-1 scenario，we assume that new revised local standards for hazardous species content in feed coal for residential uses are issued and implemented，and environmental supervision and enforcement improved.In the IPC-2 scenario，we presume that more rigorous incentive and penalty measures are implemented to restrain the coal consumption by the residential sector，new revised local hazardous content standards for coal are issued，and the best available control technology（BACT）for coal stoves is adopted to further lower the emission levels of various pollutants from residential coal combustion.These measures include the reduced use of high-sulfur coal，increased use of environmentally friendly coal stoves that produce fewer pollutant emissions and have higher combustion efficiencies，encouraging energy conservation，installation of central heating and use of clean energy alternatives.By taking these measures，the emission level from residential coal combustion will have 10%～20% reduction（MEP，2016）.

3　Results and discussion

3.1　Historical emission inventory from residential coal combustion

Emissions of primary air pollutants from residential coal stoves were calculated via the EF approach described above，as shown in Fig.3.The annual emissions of coal stoves varied with residential coal consumption.Emissions declined in 2002，2004 and 2008，but remained roughly constant in other years.These results were mainly caused by two factors：1）Due to the poor air quality，the government had to adopt measures to limit residential coal consumption.For example，the center area of Beijing promoted coal-to-electricity conversion，and the rural-urban fringe zone and rural area promoted integrated programs to centralize residential heating conversion from coal to electricity，geothermal/heat pump energy，or solar energy.These measures all helped to reduce coal consumption；2）The total population of Beijing continues to increase，with the huge influx of migrant floating individuals.The limited electricity capacity and natural gas pipes in the rural-urban fringe zone and remote rural area imply that using coal-fired stoves was much cheaper than other heating devices，leading to the increase in coal consumption.

The PM emitted by coal-fired stoves is significant，and PM2.5 accounts for about 31.1% of the total amount of PM.This proportion is lower than that of coal-fired utility boilers and industrial boilers.It implies that the combustion efficiency of coal stoves is low，which lead to the high CO emissions.Black smoke from stoves is significant and contains high contents of incomplete combustion byproducts such as EC，OC and BC.SO2 emissions are highly related to the content of sulfur in the burned coal.Since 1998，Beijing residents have been required to use low-sulfur coal to lower SO2 emissions from coal stoves.However，the coal used in rural areas is purchased from various sources with different prices，and controlling the quality of coal is quite difficult.Whereas，since the temperatures in the coal stove chamber are much lower than those in the utility or industrial boilers，resulting in less formation of thermal NOX.Due to the high content of volatiles and low combustion temperatures，VOCs emission rates are higher than those of utility or industrial boilers.In addition，coal stoves emit other harmful and toxic air pollutants，such as PAHs，PCDD/Fs，B[a]P and other pollutants that are the environment and human health hazards.

3.2　Temporal and spatial distributions of emissions from coal stoves in Beijing

Fig.4（a）shows the monthly profiles of emission variations from coal stoves based on the analysis of variation of monthly residential coal consumption.These data are derived from the statistical reports produced by local authorities.Although significant differences are present in the monthly profiles for various pollutants，a common feature of the profiles is that the emissions of air pollutants in the cold winter season are much higher than those in the other three seasons due to the increased indoor heating demand.In contrast，the lowest emissions occurred in the hot summer season，when the demand for hot water is lowest and home heating is not required，which leads to the much less coal consumption（Liu et al.，2013；Schleicher et al.，2015）.This situation of elevated emissions in winter plus the unfavorable meteorological conditions has inevitably contributed to the increasingly severe haze pollution episodes during the recent winters in Beijing（Porter et al.，2015）.

Coal combustion is the major source of SO2 emissions to the atmosphere，so we choose SO2 as a tracer to determine the relationship between monthly profiles of SO2 emissions from coal stoves and the average monthly atmospheric SO2 concentration（see Fig.4（b））.Although many factors affect the monthly concentration of air pollutants in the atmosphere，including emissions from both industrial and residential coal use，meteorological condition and so on.The monthly variation of SO2 emissions from industrial coal use was small，which contributed the basic year-round SO2 concentrations.Here，we mainly intended to show some links between monthly variation profiles of HAPs emissions from coal stoves and monthly average atmospheric SO2 concentrations.As discussed above，similar variations in the trends occur in different seasons.Both the highest SO2 emissions from coal stoves and the peak SO2 concentrations occur in winter.Hence，strengthening control of SO2 emissions from coal stoves in winter will significantly lower the annual average and peak concentrations of SO2 and other pollutants in the atmosphere of Beijing.

There are differences in coal quality used in the urban and rural areas.The coal used in the urban area was entirely anthracite.In the rural area，only 5%～15% of the coal was anthracite，and 85%～95% of the coal was bituminous.Emissions of primary air pollutants from coal-fired stoves in Beijing are predominantly located in rural and suburban areas，especially the rural-urban fringe zone，including the center of Changping District，the southeastern part of Tongzhou District and the southeastern part of Fangshan District，which are characteristically close to the main Six Ring road.These areas have convenient traffic access and relatively low housing and rental prices，which have attracted a significant migrant floating population to settle down.Due to the lack of natural gas sources and other centralized heating options，coal-fired stoves have become the most popular sources of heat in winter，causing extremely serious air pollution in the rural-urban fringe zones as a result of the low chimney height and high emission intensity.

Because urban areas adopted a series of enforcement measures to control residential coal combustion，including the establishment of the High-Polluting Fuel Ban Area，residential coal-to-electricity conversion in cultural preservation districts，electricity price allowances and other economic incentive measures，the center districts of Beijing has achieved much lower pollution levels associated with residential coal consumption than nearby surrounding areas（see Fig.5）.

3.3　Uncertainties

Several factors influence the estimation of atmospheric emissions from coal stoves in Beijing，including the reliability of activity levels and the adopted specific EFs for various pollutants.To better understand the uncertainties in our inventory，a Monte Carlo simulation is adopted to quantify the potential uncertainties of key HAPs emissions from coal stoves.By means of 8，000 rounds of repeated sampling，uncertainties in the investigated data on activity level and EFs are transmitted to the output via simulation equations.Detailed information about the input parameters and results（expressed as the lower and upper bounds of a 95% confidence interval，CI，around a central estimate）is thereby obtained.The input parameters of specific activity and emission factors，with corresponding statistical distributions are determined based on the authors’ judgment or the related published literature（Qiu et al.，2014；Tian et al.，2012；Zhao et al.，2011），as listed in Table 2.

Table 3 summarizes the uncertainties in atmospheric emission estimations of HAPs from coal stoves in 2012.

As mentioned above，the emission estimations of several pollutants，such as PM10，PM2.5，BC，EC，and OC，whose EFs are determined based on practical monitoring experiments，show relatively small uncertainties in the range of -23.1%～+25.4%.For PAHs，PCDD/Fs，B[a]P and HCl，the ranges of uncertainty are relatively large（-65.9%～+77.3%），mainly due to the lack of local EFs determined through adequate field measurements.Therefore，more field tests and surveys are necessary to better understand and quantify the emission levels of various HAPs.

3.4　Contribution of coal-fired stoves to air quality in Beijing

The Weather Research Forecast-Community Multi-Scale Air Quality（WRF-CMAQ）modeling system was used in this study to simulate the effects of residential coal combustion on atmospheric concentrations in Beijing，2012（see Fig.6）.Before the simulation，the modeling system was well evaluated by comparing the simulated 1-h mass concentrations of each pollutant（i.e.，PM10，PM2.5，SO2，and NO2）with observational data from 12 monitoring sites within the Beijing area.The evaluation protocol can be found in（Wang et al.，2010）.Table 4 shows the calculated statistical results.CMAQ reproduces well the temporal variation of each pollutant，with a correlation coefficient ranged from 0.65 to 0.70.Compared with observations，CMAQ underestimated each pollutant slightly.The range of NMB was from -0.11 to -0.24.This may be caused by the uncertainty of the emission inventory in Beijing and surrounding areas because of the incomplete source information.Furthermore，some other factors such as grid resolution，chemical mechanism and model parameterization affect the model performance as well.Generally speaking，the model performance is comparable to the literature（Simon et al.，2012），the simulation errors are acceptable，and the modeling system can be used in the following study.

The results showed that the residential coal combustion contributed relatively little to the annual air pollutant concentrations of PM10，PM2.5，SO2，NOX and CO，and that the average concentration contributions were about 4.9μg/m3，3.4μg/m3，5.4μg/m3，0.9μg/m3 and 54.1μg/m3，accounting for 4.5%，19.3%，1.7% and 3.9% of the annual average concentrations of each pollutant above，respectively.However，residential coal combustion imposed much higher contribution in winter，with PM10，PM2.5，SO2，NO2，CO average contribution concentrations of about 12.6μg/m3，8.5μg/m3，14μg/m3，1.7μg/m3 and 137.8μg/m3，which were almost 2 times that of the total annual contribution and accounted for 11.6%，27.5%，2.8% and 7.3% of concentrations of air pollutants in the heating season.In winter，residential coal combustion has a great effect on the ambient SO2 and CO concentrations due to the uncontrolled and insufficient combustion.

3.5　Future prediction and control measures

On account of the above emissions of various air pollutants from residential coal combustion and their impacts on ambient air quality，controlling residential coal combustion is quite critical to improve regional air quality in Beijing，especially in the winter heating season.In this study，to predict and analyze the future emissions of residential coal combustion，the BAU，IPC-1 and IPC-2 modeling systems were used，and the projection results were shown in Fig.7.

Under the BAU scenario，the residential coal combustion continued to follow its current trends.In the urban area，a good natural gas supply and centralized heating system were available，and increasing numbers of residents used electricity or centralized heating，resulting in decreasing trends of directly residential coal consumption.In the rural area，the use of coal-fired stoves will lead to increased coal consumption due to increases in residents’ income and higher demand for heating.And we assume the coal quality will remain at the same level in the rural area.Overall，the residential coal consumption will slightly increase with a growth rate of 0.3% under BAU scenario.Consequently，the model predicted that each air pollutant emitted by residential coal combustion would be about 1.21 times higher in 2030 than that in 2012.

Under the IPC-1 scenario，the serious air pollution in Beijing associated with coal combustion sources has been reduced through various measures.For example，power stations and industrial boilers have been gradually improved，and dust removal，de-SO2 and de-NOX control measures have been completely applied.Controlling residential coal combustion will be one of the most important measures in the future.According to the Clean Air Action Plan（BMG，2013），the use of electricity，natural gas，civil thermal，LPG，solar and other clean energy sources should be promoted through administrative and economic encouragement in areas with a good clean energy conversion condition.For areas without a clean energy conversion environment，the conversion from smoke coal to low-sulfur coal should be encouraged，and stringent quality standards and strict inspection of coal wholesale and retail market should be implemented.

According to the action schemes for high-pollution combustion zones of Beijing，6 urban districts and 80% of 10 rural downtown areas must become High-Polluting Fuel Ban Areas（BMG，2014）.The range of High-Polluting Fuel Ban Areas will be extended from Dongcheng and Xicheng District to the four urban districts（Chaoyang，Fengtai，Shijingshan and Haidian Districts）then to built-up area in suburban districts，and finally to the wide rural area.With these measures，residential coal consumption will be limited to approximately 1.44 million tons in 2030 and low-sulfur coal will be used in 55% of the rural areas.In the simulation，air pollutant emissions from residential coal combustion will be dramatically decreased，and the resulted emissions level in 2030 will be approximately 45% of that in 2012.

Under the IPC-2 scenarios，the air pollution is considered so serious that much more stringent control measures on residential coal combustion are adopted，such as difining more High-Polluting Fuel Ban Areas and stricter enforcements，to optimize the energy structure and attain the national air quality standard（GB 3095—2012）.Additionally，key rural areas have completed the clean energy conversion.Mountainous non-key rural areas find it difficult to complete the clean energy conversion；thus，many remote areas still require the use of coal-fired stoves but will be required to burn low-sulfur coal.Consequently，residential coal consumption in 2030 will be reduced to about 0.9 million tons，and the air pollutant emission levels from residential coal combustion are already significantly decreased to approximately 25% of those in 2012.

4　Conclusions

Ahousehold-based emission inventory of residential coal combustion in Beijing is developed using a bottom-up method based on an available activity database and emission factors.Although totally 16 staged air pollution control measures have been enforced since 1998 in Beijing，residential coal combustion remains，resulting in a rather high emission level.In terms of spatial distribution，residential coal combustion occurs primarily in the urban-rural integrated areas and rural areas near Six Ring road.These areas are featured with convenient traffic access and relatively low housing and rental prices and have therefore attracted a large amount of migrant floating population.However，no adequate natural gas pipes and other centralized heat sources exists.Thus，coal-fired stoves，which are cheap and easily available，are the most popular source of heating in winter，leading to heavy emissions of various air pollutants.

Regional air quality model（CMAQ）is applied to analyze the impact of residential coal combustion on the air quality in Beijing in 2012.Our analysis indicates that the contributions of residential coal combustion to air pollutants concentration in the winter heating season should be highlighted，as they represent approximately 11.6%，27.5%，2.8% and 7.3% of the total concentrations of PM10，SO2，NOX，and CO，respectively.Residential coal combustion has promoted the formation of extremely serious air pollution episodes in a great extent.

From the historical emission trends of residential coal combustion，we conclude that it is difficult to reduce the emission from residential sources merely by mandatory control measures.The best way to control air pollution from residential coal combustion is to adopt economic encouragement measures that promote the conversion from coal to clean energy sources，such as electricity and LPG，for residential heating.For example，in areas with a secure energy supply，natural gas-fired wall-mounted stoves，ground-source heat pumps，solar energy and electricity can be used；whereas in areas without a secure energy supply，low-sulfur and low-ash coal can be used to reduce the emissions of particulate matter，SO2 and other hazardous air pollutants.

■ACKNOWLEDGMENTS

This work was funded by the National Natural Science Foundation of China（21377012，21177012 and 40975061），the National Science and Technology Support Program of the Ministry of Science and Technology of China（2014BAC23B02 and 2014BAC06B05）and the Science Foundation of Beijing Municipal Research Institute of Environmental Protection（No.2014A04）.The authors also thank the MEIC team from Tsinghua University for providing the 36-km resolution emission inventory.

References

BMBS，（Beijing Municipal Bureau of Statistics），2014.Beijing Statistical Yearbook.http：//www.bjstats.gov.cn/nj/main/2014-tjnj/CH/index.htm.

BMRIEP，（Beijing Municipal Research Institute of Environmental Protection），2013.Primary air pollutats emission inventory of Beijing，2012.

BMG，（Beijing Municipal Government），2013.Clean Air Action Plan in Beijing from 2013 to 2017.http：//zhengwu.beijing.gov.cn/ghxx/qtgh/t1324558.htm.

BMG，（Beijing Municipal Government），2014.Action schemes for high-pollution combustion zones in Beijing.http：//zhengwu.beijing.gov.cn/gzdt/gggs/t1363285.htm

Chang W.，Liao H.，Xin J.，Li Z.，Li D.，Zhang X.，2015.Uncertainties in anthropogenic aerosol concentrations and direct radiative forcing induced by emission inventories in eastern China.Atmospheric Research 166，129-140.

Chen Y.，Tian C.，Feng Y.，Zhi G.，Li J.，Zhang G.，2015.Measurements of emission factors of PM2.5，OC，EC，and BC for household stoves of coal combustion in China.Atmospheric Environment 109，190-196.

Djalalova I.，Delle Monache L.，Wilczak J.，2015.PM2.5 analog forecast and Kalman filter post-processing for the Community Multiscale Air Quality（CMAQ）model.Atmospheric Environment 119，431-442.

Edwards R.D.，Smith K.R.，Zhang J.，Ma Y.，2004.Implications of changes in household stoves and fuel use in China.Energy Policy 32，395-411.

Foley K.M.，Hogrefe C.，Pouliot G.，Possiel N.，Roselle S.J.，Simon H.，Timin B.，2015.Dynamic evaluation of CMAQ part I：Separating the effects of changing emissions and changing meteorology on ozone levels between 2002 and 2005 in the eastern US.Atmospheric Environment 103，247-255.

Fu Y.，Liao H.，2012.Simulation of the interannual variations of biogenic emissions of volatile organic compounds in China：Impacts on tropospheric ozone and secondary organic aerosol.Atmospheric Environment 59，170-185.

Gan C.M.，Binkowski F.，Pleim J.，Xing J.，Wong D.，Mathur R.，Gilliam R.，2015.Assessment of the aerosol optics component of the coupled WRF-CMAQ model using CARES field campaign data and a single column model.Atmospheric Environment 115，670-682.

Gao J.，Tian H.，Cheng K.，Lu L.，Zheng M.，Wang S.，Hao J.，Wang K.，Hua S.，Zhu C.，Wang Y.，2015.The variation of chemical characteristics of PM2.5 and PM10 and formation causes during two haze pollution events in urban Beijing，China.Atmospheric Environment 107，1-8.

Ge S.，Xu Chow J.C.，Watson J.，Sheng Q.，Liu W.，Bai Z.，Zhu T.，Zhang J.，2004.Emissions of Air Pollutants from Household Stoves：Honeycomb Coal versus Coal Cake.Environmental Science & Technology 38，4612-4618.

Hogrefe C.，Pouliot G.，Wong D.，Torian A.，Roselle S.，Pleim J.，Mathur R.，2015.Annual application and evaluation of the online coupled WRF-CMAQ system over North America under AQMEII phase 2.Atmospheric Environment 115，683-694.

Huang X.，Liu Z.，Zhang J.，Wen T.，Ji D.，Wang Y.，2016.Seasonal variation and secondary formation of size-segregated aerosol water-soluble inorganic ions during pollution episodes in Beijing.Atmospheric Research 168，70-79.

Li J.，Huang X.，Yang H.，Chuai X.，Li Y.，Qu J.，Zhang Z.，2016.Situation and determinants of household carbon emissions in Northwest China.Habitat International 51，178-187.

Li L.Y.，Xie S.D.，2014.Historical variations of biogenic volatile organic compound emission inventories in China，1981-2003.Atmospheric Environment 95，185-196.

Liu X.G.，Li J.，Qu Y.，Han T.，Hou L.，Gu J.，Chen C.，Yang Y.，Liu X.，Yang T.，Zhang Y.，Tian H.，Hu M.，2013.Formation and evolution mechanism of regional haze：a case study in the megacity Beijing，China.Atmospheric Chemistry and Physics 13，4501-4514.

Lv B.，Zhang B.，Bai Y.，2016.A systematic analysis of PM2.5 in Beijing and its sources from 2000 to 2012.Atmospheric Environment 124，Part B，98-108.

MEP，（Ministry of Environmental Protection），2016.Technical guidelines for integrated abatement of residential coal-fired combustion（on trial）.http：//www.zhb.gov.cn/gkml/hbb/bgth/201603/t20160315_332883.htm

Paradiž B.，Dilara P.，Horák J.，De Santi G.，Christoph E.H.，Umlauf G.，2008.An integrated approach to assess the PCDD/Fs emissions of the coal fired stoves combining emission measurements and ambient air levels modelling.Chemosphere 73，S94-S100.

Porter W.C.，Heald C.L.，Cooley D.，Russell B.，2015.Investigating the observed sensitivities of air-quality extremes to meteorological drivers via quantile regression.Atmospheric Chemistry and Physics 15，10349-10366.

Qiu P.，Tian H.，Zhu C.，Liu K.，Gao J.，Zhou J.，2014.An elaborate high resolution emission inventory of primary air pollutants for the Central Plain Urban Agglomeration of China.Atmospheric Environment 86，93-101.

Schleicher N.J.，Schaefer J.，Blanc G.，Chen Y.，Chai F.，Cen K.，Norra S.，2015.Atmospheric particulate mercury in the megacity Beijing：spatiotemporal variations and source apportionment.Atmospheric Environment 109，251-261.

Scott A.J.，Scarrott C.，2011.Impacts of residential heating intervention measures on air quality and progress towards targets in Christchurch and Timaru，New Zealand.Atmospheric Environment 45，2972-2980.

Shen G.，Yang Y.，Wang W.，Tao S.，Zhu C.，Min Y.，Xue M.，Ding J.，Wang B.，Wang R.，Shen H.，Li W.，Wang X.，Russell A.G.，2010.Emission Factors of Particulate Matter and Elemental Carbon for Crop Residues and Coals Burned in Typical Household Stoves in China.Environmental Science & Technology 44，7157-7162.

Shi Y.，Matsunaga T.，Yamaguchi Y.，2015.High-Resolution Mapping of Biomass Burning Emissions in Three Tropical Regions.Environmental Science & Technology 49，10806-10814.

Simon H.，Baker K.R.，Phillips S.，2012.Compilation and interpretation of photochemical model performance statistics published between 2006 and 2012.Atmospheric Environment 61，124-139.

Sun Y.L.，Wang Z.F.，Du W.，Zhang Q.，Wang Q.Q.，Fu P.Q.，Pan X.L.，Li J.，Jayne J.，Worsnop D.R.，2015.Long-term real-time measurements of aerosol particle composition in Beijing，China：seasonal variations，meteorological effects，and source analysis.Atmospheric Chemistry and Physics 15，10149-10165.

Syrakov D.，Prodanova M.，Georgieva E.，Etropolska I.，Slavov K.，2016.Simulation of European air quality by WRF-CMAQ models using AQMEII-2 infrastructure.Journal of Computational and Applied Mathematics 293，232-245.

Tian H.，Gao J.，Lu L.，Zhao D.，Cheng K.，Qiu P.，2012.Temporal Trends and Spatial Variation Characteristics of Hazardous Air Pollutant Emission Inventory from Municipal Solid Waste Incineration in China.Environmental Science & Technology 46，10364-10371.

Tian S.，Pan Y.，Liu Z.，Wen T.，Wang Y.，2014.Size-resolved aerosol chemical analysis of extreme haze pollution events during early 2013 in urban Beijing，China.Journal of Hazardous Materials 279，452-460.

Wang L.，Wei Z.，Wei W.，Fu J.S.，Meng C.，Ma S.，2015a.Source apportionment of PM2.5 in top polluted cities in Hebei，China using the CMAQ model.Atmospheric Environment 122，723-736.

Wang N.，Guo H.，Jiang F.，Ling Z.H.，Wang T.，2015b.Simulation of ozone formation at different elevations in mountainous area of Hong Kong using WRF-CMAQ model.Science of The Total Environment 505，939-951.

WHO，（World Trade Organization），2005，Air quality guidelines for particulate matter，ozone，nitrogen dioxide and sulfur dioxide.http：//www.who.int/phe/health_topics/outdoorair/outdoorair_aqg/en.

Wornat M.J.，Ledesma E.B.，Sandrowitz A.K.，Roth M.J.，Dawsey S.M.，Qiao Y.-L.，Chen W.，2001.Polycyclic Aromatic Hydrocarbons Identified in Soot Extracts from Domestic Coal-Burning Stoves of Henan Province，China.Environmental Science & Technology 35，1943-1952.

Wang X.，Zhang Y.，Hu Y.，Zhou W.，Lu K.，Zhong L.，Zeng L.，Shao M.，Hu M.，Russell A.G.，2010.Process analysis and sensitivity study of regional ozone formation over the Pearl River Delta，China，during the PRIDE-PRD2004 campaign using the Community Multiscale Air Quality modeling system.Atmospheric Chemistry and Physics 10，4423-4437.

Xue Y.，Tian H.，Yan J.，Xiong C.，Pan T.，Nie L.，Wu X.，Li J.，Wang W.，Gao J.，Zhu C.，Wang K.，2016.Present and future emissions of HAPs from crematories in China.Atmospheric Environment 124，Part A，28-36.

Zhang Y.，Zhang X.，Wang L.，Zhang Q.，Duan F.，He K.，2016.Application of WRF/Chem over East Asia：Part I.Model evaluation and intercomparison with MM5/CMAQ.Atmospheric Environment 124，Part B，285-300.

Zhang Z.，Zhang X.，Gong D.，Quan W.，Zhao X.，Ma Z.，Kim S.-J.，2015.Evolution of surface O3 and PM2.5 concentrations and their relationships with meteorological conditions over the last decade in Beijing.Atmospheric Environment 108，67-75.

Zhao Y.，Nielsen C.P.，Lei Y.，McElroy M.B.，Hao J.，2011.Quantifying the uncertainties of a bottom-up emission inventory of anthropogenic atmospheric pollutants in China.Atmospheric Chemistry and Physics 11，2295-2308.

Zhi G.，Chen Y.，Feng Y.，Xiong S.，Li J.，Zhang G.，Sheng G.，Fu J.，2008.Emission Characteristics of Carbonaceous Particles from Various Residential Coal-Stoves in China.Environmental Science & Technology 42，3310-3315.