Estimation of abatement potentials and costs of air pollution emissions in China

Abstract

Understanding the air pollution emission abatement potential and associated control cost is prerequisite for designing cost efficient control policy. In this study, a linear programming algorithm model (ICET) was updated with local cost data for applications of 56 types of end-of-pipe technologies and five types of renewable energy in 10 major sectors including power plants, industry combustion, cement, steel, other industry process, domestic, transportation, solvent use, livestock, and fertilizer. The updated model was implemented for estimating abatement potential and marginal cost of multiple pollutants in China. The total maximum abatement potentials of SO₂, NO_x, primary PM_2.5, NMVOCs and NH₃ in China were estimated to be 19.2 Mt, 20.8 Mt, 9.1 Mt, 17.2 Mt and 8.6 Mt, respectively, which accounts for 89.7%, 89.9%, 94.6%, 74.0%, and 80.2% of their total emissions in 2014, respectively. The associated control cost of such reductions was estimated as 92.5, 469.7, 75.7, 449.0, and 361.8 billion CNY for in SO₂, NO_x, primary PM_2.5, NMVOCs and NH₃, respectively. Shandong, Jiangsu, Henan, Zhejiang, and Guangdong provinces exhibit large abatement potential for all pollutants. The end-of-pipe technology tends be a cost-efficiency way to control pollutions in industry process sector, while it is less cost-effective in fossil fuel-related sectors compared to the way of renewable energy application. The abatement potentials and marginal abatement cost curves developed in this study can further be used to as a crucial component in the integrated model for designing optimized cost-efficiency control policy.

1. Introduction

Air pollution is one of the biggest challenges facing China during the past two decades. Most regions in China have been suffering from the serious air pollution problem with high concentration of fine particulate matter (PM_2.5) in the world (van Donkelaar et al., 2015). Fast growth of economy and urbanization in China leads to rapid increase of anthropogenic emissions of sulfur dioxide (SO₂), nitrogen oxides (NO_x), primary PM_2.5, non-volatile organic compounds (NMVOCs), and ammonia (NH₃), resulting in degraded regional air quality and visibility (Huang et al., 2014, Wang et al., 2014, Zhao et al., 2017, Zhao et al., 2018), adverse human health impacts (van Donkelaar et al., 2015, Heo et al., 2016, Lu et al., 2016, Wang et al., 2016), economic loss (Chestnut and Mills 2005, Zhao et al., 2017, Burke et al., 2018), and global climate change over the last two decades (Bollen et al., 2009, Hill et al., 2009, Buonocore et al., 2015, Millstein et al., 2017).

To improve the air quality, the Chinese government released a series of policies and legislations to implement end-of-pipe controls and adjust the structure of energy consumption. In 2012, the first national air quality standard for PM_2.5 (GB 3095–2012) was promulgated as the Grade II Ambient Air Quality Standard of PM_2.5 is below 35 μg m⁻³. In September 2013, the Chinese government issued the Action Plan on Prevention and Control of Air Pollution (noted as “Action Plan”) aiming to reduce PM_2.5 concentrations by 10% from 2013 to 2017 nationwide. A series of control measures were taken during the Action Plan, including relocation or close down of highly polluted sources (e.g., steelmaking factories, oil refineries), implementation of advanced end-of-pipe control measures with higher removal efficiency to meet more stringent emission standards for industrial facilities, and upgrades of numerous coal-fired boilers and domestic stoves to use natural gas instead of coal (Zhao et al., 2011, Zhao et al., 2017). The effectiveness of the Action Plan is witnessed by the significant improvement in air quality. The annual averaged PM_2.5 concentrations of Beijing-Tianjin-Hebei region (JJJ), Yangtze River Delta region (YRD) and Pearl River Delta region (PRD) were decreased from 110 μg m⁻³, 70 μg m⁻³ and 48 μg m⁻³ in 2013 to 85 μg m⁻³, 55 μg m⁻³ and 34 μg m⁻³ in 2015 respectively (Wang et al., 2017). However, the cost associated with the control measures during Action Plan is considerable, which was estimated as a loss of 4.8% total GDP in JJJ in 2017 and 10.25% of that in 2020 (Minjun Shi et al., 2017).

Previous studies suggested that up to 80% of the cost associated with air pollution controls can be reduced by applying a cost-effective strategy (Amann et al., 2008). The cost-benefit analysis which has been widely applied in fields of environmental policy provided a unique framework for quantifying the cost and benefits of proposed policy actions to select the optimized strategy since the late 1960s (Pearce 1998). At present, several optimization frameworks have been developed to design least-cost strategies to control multiple precursor emissions for air pollution, including the Greenhouse Gas-Air Pollution Interactions and Synergies (GAINS) model developed by the International Institute for Applied Systems Analysis (IIASA) in Europe (Schopp et al., 1999, Amann et al., 2001, Klimont et al., 2002, Cofala et al., 2004, Amann et al., 2008, Amann et al., 2011, Klimont and Winiwarter 2011, Kanada et al., 2013, Liu et al., 2013), the Control Strategy Tool (CoST) developed by the Environmental Protection Agency (EPA) in United States (Alison Eyth et al., 2008, Loughlin et al., 2017), and AIM/Enduse model developed by the National Institute for Environmental Studies (NIES) in Japan (Hu et al., 2003, Xing et al., 2015).

In recent years, a new policy-oriented integrated scientific assessment system, the Air Benefit and Cost and Attainment Assessment System (ABaCAS) was developed by Tsinghua University, South China University of Technology and University of Tennessee and supported by U.S. EPA, aiming to address the key question whether the proposed control strategy will be cost-efficient (Xing et al., 2017). The International Control Cost Estimate Tool (ICET) module in ABaCAS was developed to calculate the cost of multi-pollutant control strategies by employing an linear programming algorithm to optimize the emissions reduction and control technologies for certain sector (Sun et al., 2014). The ICET was applied to the study of air pollution abatement in Chinese coal-fired power industry (Sun et al., 2014) as well as co-benefit analysis of energy efficiency improvement in US (Loughlin et al., 2017). However, ICET still suffers uncertainties from the lack of detailed information including the multi-pollutant control efficiencies and associated costs, particularly for sectors such as iron and steel, cement, vehicles, industry processes, as well as the end-of-pipe controls and potential applications of NMVOCs and NH₃ in China. Moreover, previous studies also found that increasing energy efficiency and applying renewable energy could significantly reduce air pollution control costs (Dai et al., 2016). For instance, about 17–35% costs for SO₂ emission reduction were reduced due to the application of renewable energy in China (Boudri et al., 2002). However, the cost of renewable energy application in China was still not implemented in ICET.

Hence, this study aims to improve ICET model with updates of local database including the cost of end-of-pipe technologies and the renewable energy application in China. The abatement potential of multiple emission sources and the control cost associated with certain emission reduction target are calculated in this study to provide a basic dataset for policy makers to design cost-effective control strategy in China.

The structure of this paper is as follows. Section 2 and 3 describes the cost model methodology, data sources and end-of-pipe control options for each pollutant by sectors. The results of abatement potential analysis and marginal abatement cost curves are given in Section 4. Section 5 highlights the conclusion.

2. Method and data

2.1. Control cost of end-of-pipe technologies

The cost of end-of-pipe control technology i for pollutant p at sector s is estimated by the following equations. First, the total cost is the sum of the capital cost, fixed operating and maintenance costs, and the fuel cost, as shown in E1.

$$T{\mathit{\text{Cost}}}_{i,p,s}=C{C}_{i,p,s}+{\mathit{\text{FOM}}}_{i,p,s}+{\mathit{\text{FUEL}}}_{i,p,s}$$ (1)

where, TCost_i,p,s is the total cost of end-of-pipe measure i for pollutant p at sector s; CC_i,p,s is the annual average capital cost of end-of-pipe measure i for pollutant p at sector s; FOM_i,p,s is the fixed operating and maintenance (FOM) costs of end-of-pipe measure i for pollutant p at sector s; FUEL_i,p,s is the fuel cost per year of end-of-pipe measure i for pollutant p at sector s. The unit is CNY/kW for power plant and CNY/t for steelmaking and cement.

Second, the annual average capital cost is estimated from the initial investment cost, the discount rate, and lifetime as shown in E2.

$${CC}_{i,p,s}={\mathit{\text{Cost}}}_{i,p,s}\frac{{\alpha (1+\alpha )}^t}{{(1+\alpha )}^t-1}$$ (2)

Where, Cost_i,p,s is the initial investment cost of end-of-pipe measure i for pollutant p at sector s; α is discount rate; t is the lifetime of end-of-pipe measure i.

In this study, a discount rate of 10% and the lifetime of 30 years were assumed. The initial investment cost and FOM cost are gathered from government documents, reports, literature research and field investigation (see detail in Table S1). The unit cost of end-of-pipe measure i for pollutant p (SO₂, NO_x, primary PM_2.5) is calculated for different types of end-of-pipe technologies in 10 major sectors (i.e., power plants, industry combustion, cement, steel, other industry process, domestic, transportation, solvent use, livestock, and fertilizer). The unit cost of most NMVOCs-related and NH₃-related sectors are directly collected from literature research or field investigation, including various end-of-pipe technologies in NMVOCs-related and NH_3-related sectors.

In order to be consistent with other pollutants, the NH₃ abatement cost in per unit of live animal which is derived from literature research in this study will be transformed to be per unit of removed NH₃ emission, by the following equation (Klimont and Winiwarter 2011).

$${uc}_{ij}=\frac{{ua}_{ij}}{n_{ij}\bullet {ef}_j}$$ (3)

where n_ij is removal efficiency of technology i for livestock category j; ef_j is the emission factor for livestock category j; ua_ij is the unit cost per live animal; and uc_ij is the unit cost per abated NH₃ emission.

2.2. Control cost of renewable energy application

In addition to the end-of-pipe controls, the measures by applying renewable energy were also considered in this study. The renewable energy, e.g., hydropower, solar power, wind, nuclear and biomass energy was considered as a replacement of fossil fuels to be applied in power generation, industry combustion and residential sectors to reduce the pollutant emissions. Similar to end-of-pipe technologies, the emission reductions and associated costs for renewable energy applications were mainly based on levelized cost published by the International Renewable Energy Agency (IRENA) in 2014 (IRENA 2015). In short, the abatement potential and associated costs of renewable energy were estimated as follows.

First, the potential reduction emissions of each renewable energy were simply assumed to be equal to the emissions from coal-fired power plant and industry combustion when the same amount of electricity or heat is produced, and the associated levelized cost of renewable energy was considered as the unit cost of each abated pollutants.

Second, the renewable energy can only replace part of the related fossil fuels, and the maximum substitution rate was determined based on the potential power generation of each renewable according to the latest China 2050 High Renewable Energy Penetration Scenario and Roadmap Study reported by Energy Research and Institute of Nation Development and Reform Commission (ERINDRC) in 2015 (ERINDRC 2015).

Third, the renewable energy application was treated as an option of control technology. The unit cost of pollutant reduction associated with the renewable energy application was calculated and implemented into the development of marginal cost curves for all pollutants.

2.3. Estimation of marginal abatement cost and abatement potential

The marginal abatement cost curves for pollutant emissions were established based on the ICET module in the ABaCAS system. First, the unit cost per ton of pollutant emission for each control technology was calculated based on the control cost and emission abatement. Second, the order of all control technologies was sorted by the unit cost per ton from low to high, implying that most cost-efficient control technology will be first to apply with its maxima application rate. Then the maxima abatement potential of each pollutant was estimated as the maxima abatement rate of emission when all available control technologies are applied.

Current control applications, as well as the unit cost, potential application rate and emission control efficiency of different control technologies are calculated as follows.

$$T{\mathit{\text{Cost}}}_{i,p,s}^r=U{C}_{i,p,s}\times \Delta {\mathit{\text{Emis}}}_{i,p,s}^r$$ (4)

$$\Delta {\mathit{\text{Emis}}}_{i,p,s}^r=\left(1-C{E}_{i,p,s}\right)\times \left(\mathit{\text{App}}{R}_{i,p,s}^r-\mathit{\text{Cur\_App}}{R}_{i,p,s}^r\right)\times {\mathit{\text{Unabated\_Emis}}}_p^{r,s}$$ (5)

$$\mathit{\text{Cur\_App}}{R}_{i,p,s}^r\le \mathit{\text{App}}{R}_{i,p,s}^r\le \mathit{\text{max\_App}}{R}_{i,p,s}$$ (6)

where, ${\mathit{\text{TCost}}t}_{i,p,s}^r$ is the total cost of technology i for pollutant p (i.e., NO_x, SO₂, NH₃, NMVOCs and primary PM_2.5) at region r; UC_i,p,s is the unit cost of technology i for pollutant p; ${\Delta \mathit{\text{Emis}}}_{i,p,s}^r$ is the emission reduction by the technology i for pollutant p at region r; CE_i,p,s is the control efficiency of technology i for pollutant p; ${\mathit{\text{AppR}}}_{i,p,s}^r$ is the control application rate of technology i for pollutant p at region r; $\mathit{\text{Cur\_App}}{R}_{i,p,s}^r$ is the current control application rate of technology i for pollutant p at region r; ${\mathit{\text{Unabated\_Emis}}}_p^{r,s}$ is the unabated emissions of pollutant p at region r in sector k where control technology i is applied; and $\text{max\_}{\mathit{\text{AppR}}}_{p,i}$ is the maxima application rate of technology i.

In this study, the data of ${\mathit{\text{Unabated\_Emis}}}_p^{r,k}$, $\mathit{\text{Cur\_App}}{R}_{i,p,s}^r$ and CE_i,p,s were derived from the study of the province-level bottom-up unit-based emission inventory developed by Tsinghua University (Wei et al., 2008, Wang et al., 2011, Wei et al., 2011, Wang et al., 2014, Zheng et al., 2018). The inventory includes detailed point source information, such as energy consumption, geographical location, capacity, boiler type and emission control technologies of SO₂, NO_x and PM_2.5 derived from the National Bureau of Statistics of China (NBS). The anthropogenic emission of NO_x, SO₂, PM_2.5, NH₃ and NMVOCs in China of 2014 were estimated to be 23.10 Mt, 21.40 Mt, 9.63 Mt, 10.71 Mt and 23.23 Mt, respectively (Zheng et al., 2018). The parameters of UC_i,p,s and $\text{max\_}{\mathit{\text{AppR}}}_{i,p,s}$ were basically referred to the ICET and GAINS-Asia model with some updates for power plants and industries from literature and surveys. The total cost of control technologies for pollutants is estimated from annualized investment, FOM costs, and fuel cost over the lifetime.

When cost-efficient measures are selected in an optimal sequence, the marginal costs (${\mathit{\text{MCost}}}_{i,p,s}^r$) are calculated by Eq 7 which provides a ranking of the available emission control measures.

$$M{\mathit{\text{Cost}}}_{i,p,s}^r=\frac{T{\mathit{\text{Cost}}}_{i,p}^r-T{\mathit{\text{Cost}}}_{i-1,p}^r}{\Delta {\mathit{\text{Emis}}}_{i,p}^r-\Delta {\mathit{\text{Emis}}}_{i-1,p}^r}$$ (7)

The marginal abatement cost and abatement potential was estimated for five pollutants including SO₂, NO_x, NMVOCs, NH₃, and primary PM_2.5, individually. For the convenience of calculation, the control technologies which can simultaneously reduce more than one pollutant will be considered in the calculation of all related pollutants. Such simplification might be correct for one pollutant when calculating individually, however, it obviously overestimates the control cost for the estimation of the total cost of simultaneous control of all pollutant because of the double counting issue.

Besides the baseline emissions in 2014 (BAL), we designed two hypothesis scenarios in this study to analyze the relative contribution from end-of-pipe measures and renewable energy application. First, the emissions are estimated with maximum application of end-of-pipe measures (noted as EOP scenario). Second, the emissions are estimated with maximum application of both end-of-pipe measures and renewable energy application (noted as REN scenario). The difference between the BAL and EOP indicates the abatement potentials of air pollutants with maximum application of end-of-pipe measures. The difference between EOP and REN indicates the abatement potentials of air pollutants with further renewable energy application. Also, the difference between BAL and REN indicates the maxima abatement potentials of air pollutants.

3. Potential control technologies in China

3.1. End-of-pipe controls

3.1.1. Power Plants

Up to 67.1% of China’s power generation is coal fired according to the inventory of power plants in 2014 (Zheng et al., 2018). For SO₂, the flue gas desulfurization (FGD) is the major control measures in China with installed capacity of 824 GW in 2014. As the share of FGD technology installed in coal-fired unit increased from 12.0% in 2005 (Wang et al., 2014) to 97.0% in 2014, the SO₂ emission from power sector has been substantially reduced by 67% during the last decade.

For NO_x controls, low-NO_x burners (LNB), selective catalytic reduction (SCR), and selective non-catalytic reduction (SNCR) were the major control technologies used in China’s coal-fired power plants, as their installed capacity was about 289 GW, 15.9 GW, and 534 GW in 2014, respectively. The share of installed denitrification facilities of SCR in coal-fired unit increased to 63.0% in 2014 from 1.1% in 2005 (Wang et al., 2014), resulting in a considerable reduction in NO_x emissions from power sector by 22.5 %.

The control technologies of primary PM_2.5 in China’s power sector include fabric filter (FF), electrostatic precipitators (ESP) and combination of FF and ESP, as their installed capacities are 9.1% (7.5 GW), 77.1% and 13.8% (11.4 GW) in the year of 2014, respectively (NBS 2015). The average removal efficiency of primary PM_2.5 is over 99.75% from coal-fired power sector in 2014, which results in limited reduction potential in primary PM_2.5 from coal-fired power sector.

For oil, gas, and biomass-fired boiler, the FOM costs are investigated from local research (Table S2), which compared the costs of four types fossil fuel boiler. The specific unit costs are simply estimated by the ratios of that compared with coal-fired boiler in this study.

Table 1 summarizes the capital cost and the FOM costs of emission control technologies in coal-fired power plant. For SO₂, the FGD actually contains limestone-based wet FGD, ammonia desulfurization, double alkali desulfurization, and magnesium oxide desulfurization (Zhang et al., 2017). The range of capital cost and FOM cost for FGD was 100 to 250 CNY/kW and 0.012 to 0.022 CNY/kWh, respectively.. For NO_x control, the SCR and SNCR technologies can be installed with LNB as LNB+SCR or LNB+SNCR. The range of costs for capital cost and FOM cost for LNB+SCR with the highest remove efficiency was 60 to 220 CNY/kW and 0.006 to 0.013 CNY/kWh. For PM control, the most efficient technology is FF+ESP with capital cost and FOM cost of 45 to 90 CNY/kW and 0.001 to 0.004 CNY/kWh, respectively.

Table 1.

Parameters of emission control technologies in coal-fired power plant sector

Pollutants	Control technology	Removal percent (%)	Capital cost (CNY/kW)	FOM cost (CNY/kWh)
SO2	FGD	85 – 95	100 – 250	0.012 – 0.022
	CFB-FGD	80 – 95	~ 198.5	0.001 – 0.01
NOx	LNB	20 – 50	10 – 40	0.0002 – 0.0006
	SCR	80 – 90	50 – 180	0.006 – 0.012
	SNCR	30 –70	30 – 40	0.005
	LNB + SNCR	35 – 70	40 – 80	0.005 – 0.006
	LNB + SCR	80 – 90	60 – 220	0.006 – 0.013
PM	FF	99	40 – 80	0.002 – 0.005
	Dry-ESP	93	50 – 100	0.0004 – 0.002
	Wet-ESP	95	30 – 70	0.0003 – 0.002
	FF + ESP	99.9	45 – 90	0.001– 0.004

FGD:flue gas desulfurization; CFB-FGD:Circulating fluidized bed- flue gas desulfurization; LNB: low-NO_x burners; SCR: selective catalytic reduction; SNCR: selective non-catalytic reduction; LNB+SNCR: combination of LNB and SNCR; FF: fabric filter; Dry-ESP: dry method of electrostatic precipitators; Wet-ESP: wet method of electrostatic precipitators; FF+ESP: combination of FF and ESP

3.1.2. Iron and steel

Emissions of SO₂, NO_x and primary PM_2.5 are generated from multiple processes in the iron and steel industry. Up to 70% SO₂ emitted from sintering process where end-of-pipe measures such as FGD, circulating fluidized bed technology (LJS-FGD) and activated carbon absorbing technology can be implemented (NBS 2015). The SCR is selected to control NO_x during the production processes. The primary PM_2.5 was emitted mostly during iron making, steel making and steel rolling processes. In this study, the capital cost and FOM cost were collected of each end-of-pipe control measures during whole process of production. The unit cost of end-of-pipe measures per pollutant abatement was calculated based on the cost and emission inventory.

Note that the activated carbon absorbing method for SO₂ was employed in a few of steelmaking plants as showed in Table 2. Compared to conventional FGD technology, activated carbon absorbing exhibits higher efficiency for its simultaneous reduction of both SO₂ and NO_x and led to a bright future to apply for further reduction. Different cost parameters are applied in different sectors even for the same measures of each pollutant. For example, the unit cost of SCR in iron and steelmaking was 24300 to 27070 CNY/t which much higher that of in other sectors, i.e., the average unit cost of SCR in power plant was 8694 CNY/t. Because there are much more procedures in the whole process than power plant.

Table 2.

Parameters of emission control technologies in iron and steel sector

Pollutant	Process	Technology	Capital cost (million CNY)	FOM cost (million CNY/year)	Reduction emission (t/year)	Unit Cost (CNY/t )	Reference
SO2	Sintering	FGD	52.87	22.65	11000	2569	Dissertation of Government Document in 2012a
		LJS-FGD	49.78	21.38	4277	6233
		Activated Carbon	340.00	68.79	8300	12633
NOx	Sintering Coking Steel making	SCR	-	-	-	24300 – 27070	Dissertation of Wu (2016)
PM2.5	Steelmaking	ESP	72.71	26.03	32000	1054	Dissertation of Government Document in 2012a
	Steel rolling	HEFF	22.43	7.20	1000	9579

^a:government document named Advanced applicable technical guide for energy saving and emission reduction in the steel industry, described by website: http://miit.gov.cn/n1146290/n4388791/c4237201/content.html;

LJS-FGD: kind of circulating fluidized bed- flue gas desulfurization; HEFF: high-efficiency fabric filter;

3.1.3. Cement

In 2012, the cement demand in China was recorded at 2184 Mt which accounts for 58% of total consumption over the world (NBS 2013). As a result, cement manufacture causes environmental impacts, particularly makes a considerable contribution to the emissions. For SO₂, the combustion of fuel with high sulfur content was the major source (Zhang et al., 2015) and FGD devices can be equipped to reduce SO₂. In 2014, about 668 denitration facilities for cement clinker production were equipped SNCR technology for reducing NO_x emissions (NBS 2013). The primary PM_2.5 emissions mainly come from the grate cooler, kiln inlet, coal mill and cement mill. The ESP and FF technologies can be implemented to capture the primary PM_2.5 during processes in cement production. The initial investment cost and FOM cost are investigated from field investigation of about 150 cement plants in this study (Table S3). Similarly, the annual average capital cost and total cost of measures implemented in cement was calculated by equation (1) and equation (2), respectively.

Referring to the Table 3, the total cost of unit cement production was 0.16 to 0.46 CNY/t for SO₂ and 0.83 to 9.14 CNY/t for NO_x. For PM_2.5 control, the end-of-pipe technologies was combined together (e.g., ESP+ESP+FF) during the entire process of cement manufacturing. The total cost of unit cement production was estimated as 10.98 to 12.99 CNY/t which is consistent with that reported by MEE in 2012, i.e., 12 to 15 CNY/t for the combination of denitration and dedusting (MEE 2012).

Table 3.

Parameters of emission control technologies in cement sector

Pollutants	Process	Technology	Removal percent (%)	Total Cost (CNY/t cement)
SO2	Coal mill	FGD	80 – 90	0.16 – 0.46
NOx	Coal mill	SNCR	55 – 75	0.83 – 6.56
		LNB+SNCR	65 – 80	1.22 – 9.14
PM2.5	Before rotary kiln	ESP	99.50 – 99.97	3.97
		ESP+FF	99.80 – 99.99	4.24
		HEFF	99.80 – 99.99	4.27
	After rotary kiln	ESP	99.5 – 99.97	5.51
		ESP+FF	99.8 – 99.99	4.93
		HEFF	99.80 – 99.99	6.66
	Coal mill	HEFF	99.80 – 99.99	2.09
	Cement mill	FF	99.00 – 99.50	3.53

FGD: flue gas desulfurization; SNCR: selective non-catalytic reduction; LNB: Low-NO_x burners LNB+SNCR: combination of LNB and SNCR; HEFF: High-efficiency fabric filter; ESP: electrostatic precipitators; FF: fabric filter; ESP+FF: combination of FF and ESP

3.1.4. Transportation sector

The government issued a series of emission standards for new vehicles in China and the implementation time is listed in Table S4. For light duty vehicle, the China I, II, III, IV, V standards came into effect nationwide in 2000, 2005, 2008, 2011 and 2017 respectively. The China VI standards will be adopted in 2020. Megacities including Beijing and Shanghai take regulating vehicle standards into force 2–3 years earlier than other provinces due to more serious air pollution problem. In general, the control measures of NO_x and VOCs from vehicle emissions in China include improvement of catalyst and fuel injection method, improvement the construction of the engine combustion chamber, update on board diagnostics (OBD) system and etc. As summarized in Table 4, the unit cost and emission abatement primarily increases with vehicle standard, which came from database in Tsinghua University, government documents, and GAINS-Asia dataset.

Table 4.

Parameters of emission control technologies in transportation.

Category	Vehicle classification	Technology	lifetime	Cost (CNY/vehicle)
				NOx	reference	VOCs	reference
Road transportation	HDT-D	China I	10	2000	Wang et al., 2014	4274	Database from GAINS
	HDT-D	China II	10	4000		12823
	HDT-D	China III	10	10909		28830
	HDT-D	China IV	10	13188		106856
	HDT-D	China V	10	20254		160283
	HDT-D	China VI	10	21557		170969
	HDT-G	China I	10	2000		19590
	HDT-G	China II	10	4000		19590
	HDT-G	China III	10	3409		19590
	LDT-D	China I	10	1000		1069
	LDT-D	China II	10	2000		1959
	LDT-D	China III	10	5000		5556
	LDT-D	China IV	10	7075		9154
	LDT-D	China V	10	9608		13731
	LDT-D	China VI	10	12899		18308
	LDT-G	China I	10	1000		1781
	LDT-G	China II	10	2000		2137
	LDT-G	China III	10	2500		5093
	LDT-G	China IV	10	3103		8050
	LDT-G	China V	10	5823		12075
	LDT-G	China VI	10	6036		16100
	HDB-D	China I	10	2000		4274
	HDB-D	China II	10	4000		12823
	HDB-D	China III	10	10909		28830
	HDB-D	China IV	10	13188		106856
	HDB-D	China V	10	20254		160283
	HDB-D	China VI	10	21577		170969
	HDB-G	China I	10	2000		19590
	HDB-G	China II	10	4000		19590
	HDB-G	China III	10	3409		19590
	HDB-CNG	China I	10	2000		19590
	HDB-CNG	China II	10	4000		19590
	HDB-CNG	China III	10	3409		19590
	HDB-LPG	China I	10	2000		19590
	HDB-LPG	China II	10	4000		19590
	HDB-LPG	China III	10	3409		19590
	LDB-D	China I	10	1000		1069
	LDB-D	China II	10	2000		1959
	LDB-D	China III	10	5000		5556
	LDB-D	China IV	10	7075		9154
	LDB-D	China V	10	9608		13731
	LDB-D	China VI	10	12899		18308
	LDB-G	China I	10	1000		1781
	LDB-G	China II	10	2000		2137
	LDB-G	China III	10	2500		5093
	LDB-G	China IV	10	3103		8050
	LDB-G	China V	10	5823		12075
	LDB-G	China VI	10	6036		16100
	CAR-D	China I	10	800		1069
	CAR-D	China II	10	1600		1959
	CAR-D	China III	10	4000		5556
	CAR-D	China IV	10	5646		9154
	CAR-D	China V	10	7686		13731
	CAR-D	China VI	10	10319		18308
	CAR-G	China I	10	800	GB 18352.5–2013 GB 18352.6–2016	1781
	CAR-G	China II	10	1600		2137
	CAR-G	China III	10	2000		5093
	CAR-G	China IV	10	2482		8050
	CAR-G	China V	10	4658		12075
	CAR-G	China VI	10	5858		16100
	CAR-CNG	China I	10	800		1781
	CAR-CNG	China II	10	1600		2137
	CAR-CNG	China III	10	2000		5093
	CAR-CNG	China IV	10	2482		8050
CAR-CNG	China V	10	4658	12075
CAR-CNG	China VI	10	5858	16100
CAR-LPG	China I	10	800	1781
CAR-LPG	China II	10	1600	2137
CAR-LPG	China III	10	2000	5093
CAR-LPG	China IV	10	2482	8050
CAR-LPG	China V	10	4658	12075
CAR-LPG	China VI	10	5858	16100
Non-road transportation	MC-2	China I	5	105	GB18176–2016; GB14622–2016
	MC-2	China II	5	210
	MC-2	China III	5	525
	MC-4	China I	5	141
	MC-4	China II	5	422
	MC-4	China III	5	704
	TRCT	China I	5	700	GB 20891–2014
	TRCT	China II	5	2100
	TRCT	China III	5	3500
	TRCT	China IV	5	9800
	TRCT	China V	5	11200
	TRCT	China VI	5	15750
	AGRC	China I	5	700
	AGRC	China II	5	2100
	AGRC	China III	5	3500
	AGRC	China IV	5	9800
	AGRC	China V	5	11200
	AGRC	China VI	5	15750
	AGRM	China I	5	700
	AGRM	China II	5	2100
	AGRM	China III	5	3500
	AGRM	China IV	5	9800
	AGRM	China V	5	11200
	AGRM	China VI	5	15750
	CNSM	China I	10	700
	CNSM	China II	10	2100
	CNSM	China III	10	3500
	CNSM	China IV	10	9800
	CNSM	China V	10	11200
	CNSM	China VI	10	15750
	RAI-D	China I	20	22850	Dataset from GAINS
	RAI-D	China II	20	44255
	RAI-D	China III	20	99505
	RAI-D	China IV	20	178058
	RAI-D	China V	20	197845
	RAI-D	China VI	20	211255
	INW	China I	20	22850
	INW	China II	20	44255
	INW	China III	20	99505
	INW	China IV	20	178058
	INW	China V	20	197845
	INW	China VI	20	211255

Note: HDT-D, heavy-duty diesel truck; LDT-D, light-duty diesel truck; HDB-D, heavy-duty diesel bus; LDT-G, light-duty gasoline truck; LDB-G, light-duty gasoline bus; CAR-G, gasoline car; HDB-LPG, heavy-duty liquefied petroleum gas bus; HDB-CNG, heavy-duty compressed natural gas; CAR-LPG, liquefied petroleum gas car; CAR-CNG, compressed natural gas car; MC-2, motorcycle with 2-stroke; MC-4, motorcycle with 4-stroke; TRCT, tractors; AGRC, agricultural construction machinery; AGRM, agricultural machinery; CNSM, construction machinery; RAI-D, railway; INW, inland water transportation

In Table 4, the parameters of emission control technologies in transportation were displayed. The stricter the standards are, the higher the unit cost of removing vehicle exhaust is. The unit cost with implementation of China VI instead of China V in transportation sector is about 1100 to 1700 CNY/vehicle and the total cost is over 25.5 billion CNY/year for light duty vehicle (China VI) announced by MEE in 2016 (MEE 2016). For NO_x, it indicated that the unit cost of diesel vehicle is much higher than that of gasoline vehicle, especially for diesel high-duty truck (HDT-D). Also, the unit cost of diesel vehicle for removing VOCs is about 4 times that of gasoline one. Meanwhile, it is more difficult and expensive to control VOCs than NO_x of vehicles. The unit cost of ship is about 10 times than that of vehicles control and facing huge challenges in controlling the emissions.

3.1.5. NMVOCs-related sector

The major anthropogenic source of NMVOCs are solvent use and fossil fuel distribution which account for 37.3% and 33.7% of total NMVOCs emission in 2010 respectively (Fu et al., 2013). The government issued national standards to control NMVOCs in 2000 and the controls on NMVOCs emission gains great attention after that (Wei et al., 2011). Nevertheless, the end-of-pipe technologies mentioned in Table 5 were still not widely used. For examples, the control measures for NMVOCs are only related with fossil fuel exploitation and distribution in China which the applicate rate takes up from 5% in 2010 to 12% in 2014 (Wang et al., 2014, Zhao et al., 2017). In addition, there are still lack of highly-efficiency measures in most industries process (Wei et al., 2008).

Table 5.

The database of end-of-pipe control measures and unit cost for reducing NMVOCs.

Sector	Control technology description	Remove efficiency (%)	Unit cost (CNY/t VOCs)
			low	high	mean
Steelmaking	Good housekeeping	0.3 – 0.6	0	800	400
Coke oven	End of pipe control measures	0.85 – 0.92	0	160	80
Refinery	Leak detection and repair program	0.37	800	10400	5600
	Covers on oil and water separators	0.097	1600	3200	2400
	Combination of the above options	0.467	2400	13600	8000
Painting	Primary measures	0.27	0	1600	800
	Primary and end of pipe measures	0.5	7.2	52000	26003
Adhesive	Primary measures	0.27	0	160	80
	Primary and end of pipe measures	0.5	2400	9600	6000
Ink	Primary measures	0.27	0	160	80
	Primary and end of pipe measures	0.5	2400	9600	6000
Synthetic material	Substitution	0.58	7.2	52000	26003
	Incineration	0.77	16.8	31200	15608
	Combination of the above options	0.903	24	83200	41612
Automobile tire	Primary measures	0.3	0	800	400
	Incineration	0.75	8000	24000	16000
	Combination of the above options	0.825	8000	24800	16400
Other rubber production	Primary measures	0.3	0	800	400
	Incineration	0.75	8000	24000	16000
	Combination of the above options	0.825	8000	24800	16400
Plant oil extraction	Activated carbon absorption	0.73	10400	17600	14000
	Schumacher type DTDC and activated carbon adsorption	0.8	15600	26400	21000
	Schumacher type DTDC and new recovery section	0.83	18720	31680	25200
Pharmacy	Primary measures and low-level end-of-pipe measures	0.73	12000	28800	20400
	Primary measures and high-level end-of-pipe measures	0.88	20000	48000	34000
Food processing	End of pipe control measures	0.9	80000	80000	80000
Papermaking	End of pipe control measures	0.9	8277	10677	9477
Oil extraction	Improved ignition system on flares	0.62	36000	44800	40400
	Alternatives and increased recovery for venting	0.9	14400	17600	16000
Synthetic polymer	Leak detection and repair program	0.6	12800	46400	29600
	Flaring	0.85	2800	2800	2800
	Add-on techniques mainly thermal and catalytic incineration	0.96	22000	55600	38800
Urban heating	Replacement of advanced coal stove	0.3	800	2400	1600
Rural heating	Replacement of advanced biomass stove (e.g. better combustion condition, catalytic stove)	0.5	14400	80000	42700
	Biomass pellet stove	0.8	15200	82400	48800
Domestic solvent use	Substitution	0.25	34400	34400	34400

In this study, the end-of-pipe measures was assumed to fully apply in chemical industry, solvent use, transport and combustion when we calculated the maximum potential reduction emissions. Based on our previous studies (Wei et al., 2008, Wei et al., 2011, Fu et al., 2013, Wang et al., 2014), the activity data including annual product, technology and pollution control facilities, as well as the unit cost of end-of-pipe technologies were collected from statistical yearbook, surveys, and dataset in GAINS-Asia model. Due to large uncertainty of control measures in coke oven, painting, adhesive, ink, pharmacy and food processing, the control technology all treated as end-of-pipe control measures and the related cost is referenced to database in GAINS-Asia model. The unit costs of NMVOCs was mainly form the database in GAINS-Asia model as summarized in Table 5. Note that the directly incineration, activated carbon adsorption, substitution, and primary measures, i.e., leak detection and repair program or good housekeeping, are mainly end-of-pipes measures for most NMVOC-related sectors. The range of unit cost is from 0 to 83200 CNY/t and the average unit cost is calculated from 9642 to 25111 CNY/t NMVOCs in different sectors, which indicates that it is more expensive than that of other pollutants (i.e., SO₂, primary PM_2.5) for management in air quality.

3.1.6. NH₃-related sector

The livestock and fertilizer application are the dominant sources for NH₃ emission, which takes up to 88% of NH₃ emissions in China for 2000 (Wang et al., 2015). Currently, there are limited NH₃ controls taken in China. In this study, we investigated all available NH₃ emission control measures, and the costs of a local pig-farm were surveyed in Table S5. The database of GAINS-Asia model was used for other NH₃ sources due to lack of local information showed in Table 6. The range of unit cost of reducing ammonia is 8905 to 66241 CNY/t in livestock sector, which is selected high efficiency technologies (≥ 80%) to calculate the maximum potential reduction emission, including animal house adaption, low ammonia application, covered outdoor storage of manure, and combination of above. For fertilizer application, the urea substitution is suitable and economically with unit cost of 4348 CNY/t.

Table 6.

The database of end-of-pipe control measures and unit cost for reducing ammonia.

Sector	Activity	Control technology description	Remove efficiency (%)	Unit cost (CNY/t)
Livestock	dairy cows	High efficiency of low ammonia application	80	47553
	other cattle	High efficiency of low ammonia application	80	33822
	sheep	High efficiency of low ammonia application	80	51800
	pigs	Combination of animal house adaption and Low ammonia application	85	66241
	poultry	Combination of Low nitrogen feed, Covered outdoor storage of manure, and Low ammonia application	97	8905
	horses
	donkey
	mule
	rabbit
Fertilizer application		Urea substitution	88	4348
Other industry process		No control	-	-
Other domestic use		No control	-	-

3.2. Renewable energy application

The potential abatement emissions and costs of renewable energy sources were summarized in Figure S1 and Table 7, respectively. To simplify the estimate of the cost of renewable energy, some hypotheses were carried out as mentioned in section 2.2. Renewable power is the essential replacement for fossil fuel. By 2050, the 86% of total power generation will be renewable power in China according to the latest China 2050 High Renewable Energy Penetration Scenario and Roadmap Study reported by Energy Research and Institute of Nation Development and Reform Commission (ERINDRC) in 2015 (ERINDRC 2015). The wind, solar, hydropower, biomass and geothermal energy could meet more than 60% of primary energy demand. According to this report, the potential application of renewable energy was estimated. By 2050, the coal power generation will drop to below 7% and the wind, hydropower, solar, nuclear power, natural gas, and other energy (biomass, geothermal) could replace 35.2%, 14.4%, 28.4%, 4.3%, 3.1% and 7.9% of coal-fired power plant, respectively. Furthermore, about 60% of coal used in industry combustion, residential heating and lighting could replace by solar energy in 2050. The unit cost of wind, hydropower, solar, natural gas are 0.40, 0.43, 0.42, and 0.47 CNY/kWh respectively that is cheaper than that of coal-fired power generation with 0.53 CNY/kWh.

Table 7.

The potential application of renewable energy sources in China

Renewable energy	Renewable technology	Sector replaced	Fuel replaced	Substitution ratio (%)	Unit Costa CNY/kWh
Wind	Wind turbine	Power	Hard coal	35.2	0.40
Hydro power	Hydro power station	Power	Hard coal	14.4	0.43
Nuclear power	Nuclear power station	power	Hard coal	4.3	0.58
Natural gas	Power plant	power	Hard coal	3.1	0.47
Others				7.9	0.6
Solar	Flat plate collectors	Power	Hard coal	28.4	0.42
	Flat plate collectors	Industry combustion	Hard coal	60	1.12
	Photovoltaic cells	Residential	Electricity	60	0.42

^aUnit Cost: cost value is from the leveling cost of reports by International Renewable Energy Agency (IRENA) and calculated by exchange rate conversion of 6.8.

4. Results

4.1. Estimation of emission abatement potential in China

To analyze the abatement potential of multiple pollutants, the EOP scenario was designed assuming that the application rate of end-of-pipe control technologies reach their maxima. Figure 1 presents the changes of pollutants by sectors under three scenarios.

Figure 1.

The potential analysis of emission reduction from sectors

The emission of SO₂, NO_x, primary PM_2.5, NMVOCs and NH₃ will decrease by 69.2%, 56.1%, 53.8%, 56.6% and 80.3%, respectively under EOP scenario. Under REN scenario by applying renewable energy, the emission of SO₂, NO_x, primary PM_2.5, and NMVOCs will be 89.7%, 89.9%, 94.6%, and 74.0% lower than those in the BAL scenario. The total abatement potential of SO₂, NO_x, primary PM_2.5, NMVOCs and NH₃ in 2014 is estimated as 19.2 Mt, 20.8 Mt, 9.1 Mt, 17.2 Mt and 8.6 Mt, respectively.

For the end-of-pipe control measures, the industry sector, including cement, iron and steel and other industry processes, has the largest abatement nationwide, with SO₂, NO_x, primary PM_2.5 and NMVOCs reduced by 31.7%, 32.5%, 72.5% and 36.5%, respectively. In particular, for cement and steelmaking sectors, the application of advanced technology including SNCR for NO_x controls and high efficiency deduster (i.e., HEFF, ESP+HEFF) for PM controls in cement manufacture, as well as the activated carbon adsorption technology for collaborative reduction of SO₂ and NO_x during iron and steel making, would bring substantial emission reductions. The application of EOP in industry combustion and power plant can reduce the SO₂ emissions by 39.5% and 23.3%, and reduce PM_2.5 emissions by 16.8% and 9.1%, respectively. The abatement potential of NOx largely comes from EOP controls in power plant (26.7%) and transportation (20.7%). Petroleum industry and solvent use exhibit largest abatement potentials for NMVOCs with 32.5% and 39.2% emission reduction, respectively. The livestock sector will reach the NH₃ abatement by 62.2% due to the application of large scale raising and breeding farm.

Table 8 summaries the abatement potential and the total cost of multi-pollutants in each province in China. In this study, different cost values of the same end-of-pipe measures are investigated due to different economy level and the total cost calculated by the average value. Shandong province exhibits the largest abatement potential with SO₂, NO_x, primary PM_2.5 and NMVOCs reduced by 1993.4 kt, 2268.8 kt, 798.0 kt and 1667.4 kt, respectively. Furthermore, Henan province as one of major agricultural province exhibits the largest abatement potential of NH₃ with 748.2 kt, followed by Sichuan province and Shandong province. In JJJ region, the abatement ratio of SO₂, NO_x, primary PM_2.5, NMVOCs and NH₃ was 85.6%, 89.8%, 94.4%, 71.8%, and 78.1%, respectively. In YRD region, the abatement ratio of SO₂, NO_x, primary PM_2.5, NMVOCs and NH₃ was 93.4%, 90.6%, 96.9%, 67.3%, and 79.2%, respectively. In PRD region, the ratio of that was 92.9%, 88.2%, 96.5%, 71.1% and 81.0%, respectively. Compared with JJJ and PRD region, the YRD region has the largest potential abatement emissions of SO₂, NO_x, and primary PM_2.5, which indicated that further reduction needed to be considered in this region.

Table 8.

Potential for reducing emissions and total cost of multi-pollutants with respect to 2014.

Pollutants		SO2						NOx						PM2.5						NMVOCs						NH3c
	MPRa kt (%)		Total Costb			MPR kt (%)			Total Cost			MPR kt (%)			Total Cost			MPR kt (%)			Total Cost			MPR kt (%)			Total Cost
			low	high	mean				low	high	mean				low	high	mean				low	high	mean				low	high	mean
Beijing	55.8	(79.4)	0.3	0.4	0.4	173.3	(89.7)		8.9	9.4	9.1	34.5	(85.7)		0.2	0.2	0.2	233.6	(68.7)		4.6	9.5	7.0	26.0	(63.1)		0.0	1.4	1.4
Tianjin	212.0	(88.4)	1.0	1.5	1.2	264.4	(91.1)		6.6	7.4	7.0	88.2	(94.6)		0.5	0.7	0.6	226.5	(69.1)		3.4	8.0	5.7	37.3	(73.2)		0.0	1.8	1.8
Hebei	729.1	(85.3)	3.1	4.3	3.7	1084.3	(89.5)		31.6	34.4	33.0	661.7	(94.9)		3.6	5.2	4.4	1025.0	(73.2)		17.1	34.7	25.9	461.3	(77.0)		0.0	18.4	18.4
Shanxi	833.6	(87.9)	3.8	5.1	4.4	776.9	(91.2)		15.7	17.9	16.8	453.7	(94.4)		1.9	2.6	2.3	522.1	(79.9)		6.6	12.4	9.5	111.4	(62.5)		0.0	4.5	4.5
Neimeng	945.1	(85.8)	4.8	6.1	5.5	1035.3	(91.7)		15.8	18.8	17.3	347.3	(87.4)		2.5	3.1	2.8	459.2	(76.2)		6.4	13.3	9.8	312.1	(77.3)		0.0	9.9	9.9
Liaoning	754.2	(89.9)	2.7	5.6	4.1	914.1	(91.2)		17.4	21.0	19.2	377.3	(96.6)		2.3	3.2	2.7	731.4	(70.3)		8.5	24.7	16.6	288.0	(80.1)		0.0	13.5	13.5
Jilin	324.8	(90.5)	1.2	1.9	1.6	500.1	(91.3)		9.3	11.0	10.2	248.2	(95.8)		1.5	1.8	1.6	331.2	(76.2)		5.5	13.0	9.2	233.7	(80.3)		0.0	8.9	8.9
Heilongjiang	255.0	(90.0)	1.1	1.9	1.5	585.5	(91.6)		11.2	13.1	12.2	287.1	(93.2)		2.3	2.4	2.3	418.3	(78.4)		6.7	17.0	11.8	273.3	(78.8)		0.0	10.8	10.8
Shanghai	453.5	(89.5)	1.6	4.3	2.9	283.7	(91.7)		6.9	7.9	7.4	77.6	(93.7)		0.3	0.7	0.5	438.5	(62.8)		3.2	14.7	8.9	19.2	(63.1)		0.0	1.1	1.1
Jiangsu	900.6	(93.3)	3.6	5.0	4.3	1170.1	(90.1)		28.0	30.7	29.4	461.8	(97.4)		4.2	5.3	4.7	1371.9	(70.2)		19.6	52.8	36.2	403.7	(78.1)		0.0	13.9	13.9
Zhejiang	1079.4	(95.2)	3.9	5.1	4.5	835.4	(91.0)		20.1	21.9	21.0	226.6	(96.9)		2.2	2.6	2.4	1114.5	(65.7)		12.9	40.0	26.4	148.9	(75.7)		0.0	7.3	7.3
Anhui	529.6	(91.8)	1.6	2.7	2.2	946.4	(91.6)		25.4	27.7	26.6	431.9	(97.7)		3.9	4.5	4.2	784.3	(82.2)		13.7	33.8	23.8	329.0	(77.7)		0.0	13.9	13.9
Fujian	448.1	(93.1)	1.6	2.3	1.9	622.0	(86.8)		9.7	12.0	10.8	170.7	(96.0)		1.6	1.9	1.7	473.2	(69.7)		6.0	16.7	11.3	178.3	(79.7)		0.0	8.6	8.6
Jiangxi	344.4	(90.8)	1.0	1.6	1.3	471.9	(86.8)		9.3	10.6	10.0	222.0	(96.6)		1.9	2.4	2.2	354.1	(75.6)		6.3	13.3	9.8	242.3	(80.0)		0.0	14.4	14.4
Shandong	1993.4	(90.4)	7.8	10.8	9.3	2168.8	(91.3)		41.5	48.4	45.0	798.0	(95.3)		4.7	5.9	5.3	1667.4	(70.7)		26.5	60.3	43.4	579.2	(76.7)		0.0	24.4	24.4
Henan	837.9	(90.0)	3.0	4.2	3.6	1458.2	(90.8)		32.8	36.6	34.7	660.6	(95.4)		4.4	5.2	4.8	1033.6	(77.7)		18.9	40.5	29.7	748.2	(78.9)		0.0	30.7	30.7
Hubei	971.8	(87.4)	3.6	6.0	4.8	858.3	(87.8)		17.6	20.6	19.1	470.0	(92.8)		3.4	4.0	3.7	710.7	(76.9)		10.0	25.7	17.9	440.5	(76.7)		0.0	19.9	19.9
Hunan	683.6	(87.5)	2.5	3.7	3.1	667.6	(87.3)		14.1	16.1	15.1	325.1	(91.1)		2.9	3.4	3.1	512.6	(74.0)		8.6	19.0	13.8	471.6	(81.4)		0.0	26.9	26.9
Guangdong	885.4	(92.9)	3.5	5.1	4.3	1253.5	(88.2)		27.2	31.2	29.2	357.4	(96.5)		3.3	3.8	3.5	1182.8	(71.1)		18.8	43.7	31.2	350.6	(78.5)		0.0	16.6	16.6
Guangxi	571.4	(91.3)	2.0	2.8	2.4	493.5	(87.7)		10.0	11.5	10.7	319.6	(96.9)		2.6	3.0	2.8	506.0	(77.5)		11.1	22.4	16.8	322.3	(81.0)		0.0	16.0	16.0
Hainan	71.7	(95.2)	0.2	0.3	0.3	97.9	(89.3)		1.7	2.0	1.8	28.6	(96.7)		0.4	0.4	0.4	96.0	(70.7)		1.4	3.7	2.5	56.8	(77.1)		0.0	2.7	2.7
Chongqing	816.4	(87.3)	3.4	4.6	4.0	393.1	(88.0)		8.9	10.1	9.5	186.2	(93.4)		1.7	1.9	1.8	311.0	(79.3)		4.7	12.5	8.6	203.7	(78.7)		0.0	9.6	9.6
Sichuan	1388.6	(90.1)	5.1	7.6	6.3	848.4	(87.4)		16.3	19.3	17.8	522.6	(97.4)		4.8	5.4	5.1	1002.0	(83.2)		17.0	45.1	31.0	714.5	(80.2)		0.0	35.1	35.1
Guizhou	856.7	(84.1)	3.9	5.0	4.4	500.1	(87.3)		7.6	9.0	8.3	313.9	(87.1)		3.0	3.3	3.2	301.5	(84.2)		4.7	9.7	7.2	202.0	(72.9)		0.0	9.2	9.2
Yunnan	375.9	(87.9)	1.0	1.7	1.4	438.8	(87.9)		9.4	10.5	9.9	240.3	(93.0)		2.2	2.6	2.4	334.3	(80.5)		6.7	12.3	9.5	393.8	(80.2)		0.0	17.3	17.3
Xizang	3.3	(92.8)	0.0	0.0	0.0	22.1	(94.1)		0.8	0.8	0.8	6.0	(98.1)		0.1	0.1	0.1	13.0	(88.2)		0.5	0.7	0.6	70.9	(78.9)		0.0	2.6	2.6
Shaanxi	622.2	(87.9)	2.9	3.5	3.2	592.0	(89.7)		11.5	13.0	12.3	301.1	(95.0)		2.5	2.9	2.7	430.4	(78.5)		6.0	13.4	9.7	210.2	(76.3)		0.0	6.4	6.4
Gansu	247.4	(88.8)	0.9	1.5	1.2	320.6	(88.9)		8.6	9.3	8.9	166.7	(93.8)		1.4	1.7	1.5	206.9	(80.7)		3.8	7.5	5.7	158.0	(76.8)		0.0	5.6	5.6
Qinghai	42.6	(88.5)	0.1	0.3	0.2	97.7	(88.5)		2.1	2.4	2.2	43.9	(91.9)		0.4	0.4	0.4	46.2	(81.0)		0.9	1.6	1.3	65.6	(76.8)		0.0	2.6	2.6
Ningxia	238.0	(93.9)	1.1	1.4	1.2	190.4	(91.5)		3.5	4.0	3.8	80.1	(97.2)		0.4	0.6	0.5	71.3	(79.7)		1.3	2.2	1.7	46.0	(74.7)		0.0	1.2	1.2
Xinjiang	721.3	(93.9)	2.7	3.9	3.3	696.3	(93.5)		9.6	11.8	10.7	209.1	(95.5)		1.4	1.8	1.6	272.5	(76.6)		4.2	8.5	6.3	260.7	(76.8)		0.0	6.5	6.5
Total	19192.9	(89.7)	75.0	110.0	92.5	20760.5	(89.9)		439.0	500.4	469.7	9117.6	(94.6)		68.3	83.0	75.7	17181.9	(74.0)		265.5	632.4	449.0	8359.3	(78.0)		0.0	361.8	361.8

^aMPR: Maximum potential reduction emissions, unit: kt; (%) represents the reduction ratio to that in 2014.

^bTotal Cost: Calculated by average cost and the range of cost is in supplementary information, unit: billion;

^cNH₃: the low cost of ammonia is zero for uncontrolled in China at present.

Identifying the key province and sector which exhibits the largest potential of air pollutant is useful for policymaker to design effective control policy. From our analysis, the provinces of Shandong, Hebei, Henan and Sichuan show the largest abatement potential for SO₂, NO_x and primary PM_2.5. For NMVOCs controls, the regions of YRD have the substantial abatement potential. Compared to power plant and industry combustion sector, the cement, steelmaking and other industry process (e.g., oil refining, glassmaking and chemical industry) exhibits larger abatement potential to reduce emissions.

4.2. Estimation of marginal abatement cost curves

Figure 2 displays the total marginal abatement cost of five pollutants in China. Both end-of-pipe measures and energy structure adjustment were applied in the development of marginal abatement cost curves. In summary, the total cost of five major air pollutants is up to 1448.6 billion CNY, in which 801.5 billion CNY for EOP scenario (see detail in Table S6). Under REN scenario, the abatement cost of NO_x and NMVOCs is much higher than SO₂ and PM_2.5, about 5 to 6 times high, which is up to 469.7 billion and 449.0 billion CNY compared with SO₂ of 92.5 billion, PM_2.5 of 75.7 billion CNY and NH₃ of 361.8 billion CNY, respectively. The total cost for SO₂, NO_x and primary PM_2.5, and NMVOCs under REN scenario is 2.2, 4.8, 1.8, 1.8 times higher than that of EOP scenario, respectively. High NO_x and NMVOCs cost is associated with the lower removal percent and utilization rates with higher marginal abatement cost for denitrification and NMVOCs extraction.

Figure 2.

The marginal abatement cost of multi-pollutants in China

For SO₂ controls, the marginal abatement cost curve increases sharply when the reduction ratio reaches 10% and 83%, respectively. The low efficiency end-of-pipe measures in power plant, including turning off some power plants with installed capacity less than 100 MW, were replaced by high-efficiency FGD technology when the reduction ration reach 10%. After that of 83%, the raise of marginal abatement cost curve depended on the higher unit cost of FGD by using nature gas and liquefied petroleum gas (LPG) of boilers and domestic stoves. Furthermore, the application of energy structure adjustment by replacing fossil fuel with wind or solar energy could further reduce SO₂ emission, which indicated the renewables may be a cost-effective solution in the future.

For NO_x, three turning points appeared as the rising of the marginal abatement cost curve. When the reduction ratio is less than 58.0%, the LNB+SCR technologies and energy structure adjustment were mainly applied in power plant, industry combustion, and domestic. After that, the marginal cost increases sharply with higher cost of activated carbon absorbing method used in sintering process or the SCR technology in gas-fired or oil-fired boilers and domestic stoves. Then, the large-scale upgrade the quality of gasoline product and promote new energy electric vehicle policy, which resulted in the further abatement of NO_x after the reduction ration of 77.1%.

For primary PM_2.5, when the reduction ratio was larger than 29.3%, the substitution of renewables in coal-fired power plant were more expensive than conventional technologies (i.e., ESP, FF, ESP+FF) as depicted in the cost curve of PM_2.5. The dedusting in gas-fired or oil-fired boilers was difficultly and costly of the reduction ratio over 90.5%.

For NMVOCs, the production process of synthetic material, painting, chemical industry, plant oil extraction and solvent use of NMVOCs showed large potential abatement and marginal cost with the reduction ratio of 42.6% to 62.0%. When the reduction ratio was larger than 62.0%, the marginal cost was up to 267.3 billion because of implementation of new energy electric vehicle policy. The urea substitution was used in fertilizer application that the reduction ratio under 38.4% is costly for removing ammonia. Meanwhile it demonstrated there exit larger potential abatement and marginal cost of NH₃ in livestock if the reduction ratio exceeded 40%.

Hence, the further reduction of air pollutants should depend on the adjustment of energy structure in our country, e.g., increase of renewable energy generation proportion, closure of high-emission plants, which cause significant social marginal benefit.

Figure 3 shows the abatement potential emissions and cost in different sectors. Industry combustion exhibits the most abatement potential of SO₂ with 35.0% in all sectors, which also has the highest abatement cost of 34.4% (31.9 billion CNY). For SO₂, the power plant and domestic sectors still exit the larger abatement potential and higher cost. However, the industry process sector, including cement, iron and steel and other industry process, not only has the larger abatement potential (24.5%) but has the lower abatement cost which account for 14.9 billion CNY (16.2%). Compared to power plant and industry combustion, the controls of cement, steelmaking and other industry process are more cost-effective. Similarly, the potential reduction emission of NO_x in industry process was 4209 kt (20.3% in total potential abatement emission) with cost of 35.2 billion (7.5% in total cost). About 69.5% abatement cost of NO_x is from transportation sector due to the high technology cost and new energy vehicles. For primary PM_2.5, the most economically abatement sector is other industry process with 0.1% abatement cost, i.e., lime production, building ceramic production, glass and brick production, which has the abatement potential with 1624 kt. With the largest abatement potential of PM_2.5 (3388 kt) in domestic, the total cost was 31.1 billion CNY. For NMVOCs control, solvent use, other industry process, domestic and transportation sectors have the higher abatement potential and cost resulted from disordered management of control and high investment costs. Up to 25.6% NMVOCs could be reduced in other industry process sector, e.g., car manufacturing, vehicle refinishing, and it has the 18.0% of abatement cost (80.9 billion CNY). According to research (He et al., 2018), electric vehicles is a cost-effective way to alleviate air pollution from a life cycle perspective, which point out electric vehicles could reduce VOCs and NO_x emission by 1770 to 2354 g/year per vehicle and 930 to 1354 g/year per vehicle, respectively. Especially, the reduction potential of VOCs emissions for vehicle material cycle mainly focus on vehicle fluids (i.e., windshield washer fluid) and vehicle manufacturing (notably the painting process). The maximum potential reduction of NH₃ in livestock and fertilizer application were 5352 kt and 3254 kt with the total cost of 347.7 billion and 14.1 billion, respectively. The unit cost of NH₃ in livestock was the highest in all sectors which indicated it was more difficult and expensive for reducing NH₃ in livestock.

Figure 3.

The abatement potential and total cost by sector

4.3. Relationship between abatement cost and local economy

We also investigated the relationship of unit abatement costs and local economy at provincial level, as displayed in Figure 4. The R was the ratio of GDP at each province to the nationwide in 2014. The larger the R value, the higher the ratio of GDP. The slope represents the unit cost of abated emissions (purple line represents the national level). Generally, the total abatement cost increases along with the total abatement potential. However, differences in slope are still noticeable among provinces.

Figure 4.

The unit costs of abated air pollutants at provincial level with local economy.

For SO₂, two kinds of regions exhibit cost-effective reduction emissions, which one represented high GDP with intensive industries (i.e., Zhejiang, Shandong) and another represented lower GDP with high sulfur content of coal for reduction SO₂ emissions (i.e., Sichuan). The better quality of coal with low sulfur constant results in the higher unit cost of abatement emissions in Inner Mongolia which are the main coal mining and renewable energy supply areas as explained in previous studies (Dong et al., 2015). Meanwhile, the less cost-effective regions were Beijing, Tianjin, and Shanghai, which showed relatively lower potential reduction. That is because the end-of-pipe measures has been fully implement in those developed eastern regions and the unit cost raised for further reduction which highly depended on renewable energy application.

For NO_x, the cost-effective reduction regions were Inner Mongolia, Xinjiang, and Guizhou province. The least cost-effective regions are located in the provinces with high potential abatement, high population density and vehicle population, including Beijing, Shandong, Henan, Hebei, Guangdong, Jiangsu, Anhui, and Zhejiang.

For primary PM_2.5, Shandong, Henan, and Hebei provinces exhibit huge potential abatement and high cost to control. The most cost-effective region was Shanxi, while the least cost-effective regions included Hainan, Xizang, and Qinghai.

For NMVOCs, there was higher potential abatement and total cost in developed areas, such as Shandong, Jiangsu, Guangdong, Sichuan, Henan, and Zhejiang province.

For NH₃, the potential abatement regions included some major agricultural provinces, such as Sichuan, Henan, Hunan, and Shandong. In summary, it is important to pay more attention to regional disparity with different economic levels and energy structures for policy making to improve the whole air quality.

5. Conclusion

This study investigated the abatement potential and cost of multi-pollutants in China, by implementing local dataset which reflect the fact of actual operating conditions, including investment cost, fixed operating and maintenance cost, removal percent and application rate of end-of-pipe control technologies.

The result shows that the total maximum abatement potentials of SO₂, NO_x, primary PM_2.5, NMVOCs and NH₃ in China are 19.2 Mt, 20.8 Mt, 9.1 Mt, 17.2 Mt and 8.6 Mt, respectively, based on the emission level of 2014. For the end-of-pipe control measures, the industry sector, including cement, iron and steel and other industry process (i.e., oil refining, glassmaking and chemical industry), still has the largest abatement nationwide. The SO₂, NO_x, primary PM_2.5 and NMVOCs can be reduced by 31.7%, 32.5%, 72.5% and 36.5%, respectively after the application of end-of-pipe control measures. Compared to other regions, Shandong province has the most abatement potential with SO₂, NO_x, PM_2.5 and NMVOCs accounting for 1993.4 kt, 2168.8 kt, 798.0 kt and 1667.4 kt respectively, followed by Sichuan, Hebei, and Henan provinces.

Application of renewable energy can further reduce the pollutants but with much more cost from end-of-pipe controls. Results suggest that the emission of SO₂, NO_x, PM_2.5, and NMVOCs will be further decreased by 89.7%, 89.9%, 94.6%, and 74.0% under REN scenario based on 2014 emission level. The marginal abatement cost of air pollutants in China, including applications of both end-of-pipe measures and renewable energy application, is 92.5 billion of SO₂, 469.7 billion of NO_x, 75.7 billion of primary PM_2.5, 449.0 billion of NMVOCs, and 361.8 billion of NH₃, respectively. Comparing the limitation of conventional EOP technologies, the application of renewable energy to replace fossil fuel could result in further reduction of pollutants, which illustrated the renewable energy application can be another option besides end-of-pipe measures in the future.

Different control measures are suggested to be implemented based on local industry structure and economy level. For SO₂ and NO_x, the end-of-pipe technologies employed in industry process (i.e., lime production, building ceramic production, glass and brick production) were more cost-effective than that of in fossil fuel sectors (i.e., power plant, industry combustion, domestic). The application of renewable energy could significantly increase the marginal abatement cost for further reduction especially for SO₂ and NO_x. For PM_2.5, the total cost in cement was still larger than other sectors, suggesting implement more stringent end-of-pipe controls. However, the end-of-pipe technology, as a matter of priority, still can be a cost-efficiency way to control NMVOCs and NH₃ which have huge abatement potential in China.

The discrepancy in the economic level among provinces are also recommended to be taken into consideration when designing control policy. Provinces with higher potential abatement emissions and total abatement costs are those developed regions with higher GDP, such as Shandong, Jiangsu, Henan, Zhejiang, and Guangdong. However, the least-effective regions with high GDP, including Beijing, Shanghai, and Tianjin, have limited potential reduction due to the high unit abated cost.

In general, this study gathered most available local information about marginal abatement cost, the application rate of control measures, and further calculated the corresponding emission abatement and associated costs in China. These abatement potential analysis and costs assessment will contribute to the analysis with integrated assessment model (e.g., ABaCAS) to design cost-effective control strategies for policymakers to improve air quality. The marginal abatement cost curves developed in this study can also be used by combining with air quality model simulation, to further evaluate the cost-effectiveness of abatement policy designed to achieve certain ambient air quality target.