Environmental Beneﬁts of Ultra-Low Emission (ULE) Technology Applied in China

: Seven scenarios were designed to study the national environmental beneﬁts of ULE in coal-ﬁred power plants (CPPs), ULE in industrial coal burning (ICB) and NH 3 emission reduction by using the GEOS-Chem model. The results showed that although the CPPs have achieved the ULE transformation target, the PM 2.5 concentration across the country has decreased by 4.8% (1.4 µ g/m 3 ). Due to the complex non-linear chemical competition mechanism among nitrate and sulfate, the average concentration of nitrate in the country has increased by 1.5% (0.1 µ g/m 3 ), which has reduced the environmental beneﬁts of the power plant emission reduction. If the ULE technology is applied to the ICB to further reduce NO x and SO 2 , although the PM 2.5 concentration can be reduced by 10.1% (2.9 µ g/m 3 ), the concentration of nitrate will increase by 2.7% (0.2 µ g/m 3 ). Based on the CPPs-ULE, NH 3 emissions reduced by 30% and 50% can signiﬁcantly reduce the concentration of ammonium and nitrate, so that the PM 2.5 concentration is decreased by 11.5% (3.3 µ g/m 3 ) and 16.5% (4.7 µ g/m 3 ). Similarly, based on the CPPs-ICB-ULE, NH 3 emissions can be reduced by 30% and 50% and the PM 2.5 concentration reduced by 15.6% (4.4 µ g/m 3 ) and 20.3% (5.8 µ g/m 3 ). The CPPs and ICB use the ULE technology to reduce NO x and SO 2 , thereby reducing the concentration of ammonium and sulfate, causing the PM 2.5 concentration to decline, and NH 3 reduction is mainly achieved through reducing the concentration of ammonium and nitrate to reduce the concentration of PM 2.5 . In order to better reduce the concentration of PM 2.5 , NO x , SO 2 and NH 3 emission reduction control measures should be comprehensively considered in different regions of China. By comprehensively considering the economic cost and environmental beneﬁts of ULE in ICB and NH 3 emission reduction, an optimal haze control scheme can be determined.


Introduction
Haze pollution is a complex phenomenon, resulting from primary emissions from multiple sources interacting with meteorology and atmospheric photochemistry. Over the past decade, China has experienced severe haze pollution, marked by fine particulate matter (PM), especially in the Beijing-Tianjin-Hebei (BTH) regions, according to the air quality status reports released by China's Ministry of Environmental Protection [1]. Fine particulate matter smaller than 2.5 µg (PM 2.5 ) is particularly troublesome and has been implicated in a range of adverse health outcomes [2,3]. The annual mean PM 2.5 concentration in the BTH The environmental benefits of using GEOS-Chem to quantify the application of ULE in CPPs and ICB are based on the data source shown in Table 1. Firstly, an emission inventory of NO x , SO 2 and PM as air pollutants after ULE technology is applied in CPPs and ICB in China, based on reasonable assumptions, was established. Then, the environmental benefits brought about by the application of ULE technology in CPPs and ICB in China were quantified by using the GEOS-Chem model. Finally, some policy suggestions for achieving greater environmental benefits in the future were put forward.  [20] Coal consumption data Raw coal consumption in power plants and industries in different provinces in 2015 (detailed classification of the power and industrial sectors is listed in Table S1).

China Energy Statistics
Yearbook [21] Multi-Resolution Emission Inventory for China (MEIC) The MEIC data on spatial distribution and seasonal variation of pollutants as well as total NO x , SO 2 , PM emissions from power plants and industrial sectors were used. MEIC [7,22] Greenhouse Gas and Air Pollution Interactions and Synergies (GAINS) The GAINS inventory provides provincial-level emission data for 11 detailed sectors (details are listed in Table S2). The GAINS data on pollutant emissions from different fuel types in the power sector and industrial combustion and industrial processes and production in the industrial sector were used.
GAINS [23,24] In order to establish the emissions inventory of air pollutants by using ULE technology in CPPs, the emission factors of air pollutants after ULE technology are used in CPPs were ensured to be consistent with Liu et al.'s research results [20], which updated the emission factors of ULE technology applied in CPPs in China, and according to the national energy statistics yearbook [21], to obtain the coal consumption data of different provinces. Only NO x , SO 2 and PM pollutant emissions were considered, when estimating the emissions after the ULE technology was applied to the CPPs in China, given the limited data availability for the emission factor of CPPs that install ULE facilities.
The application of GEOS-Chem to quantify the environmental benefits in different scenarios is based on MEIC emission data and the corresponding spatial distribution. In order to quantify the environmental benefits brought about by the application of ULE technology in the CPPs and ICB, the total emissions from the power and industry sectors provided by MEIC were used. However, the total emissions from the power and industry sectors include not only coal combustion emissions but also other emissions (oil, gas and others); MEIC does not provide these detailed data. Due to the data limitations, the proportion of coal combustion and non-coal combustion obtained from GAINS was used. Moreover, we assumed that the proportion of emissions from coal combustion and non-coal combustion in the power and industry sectors on the MEIC was the same as the proportion of emissions from coal combustion and non-coal combustion in the power and industry sectors on the GAINS.

Simulation Scenario Design 2.2.1. Emissions Calculation
The emission of each species (NO x , SO 2 and PM) from CPPs after applying the ULE technology can be calculated as: Here, EF ULE is the emission factor after the ULE technology is applied in the CPP; W S is the standard coal consumption in the CPP. Since raw coal consumption is recorded in the yearbook, the standard coal consumption was converted: Here, Q S is the lower heating value for standard coal (29,307.6 J/kg), Q R is the lower heating value for raw coal (defined by the type of coal), and W R is the raw coal consumption by mass.
Here, P CM , P CG is the proportion of pollutants emitted by CPPs in the total emissions in MEIC and GAINS, respectively; E CG is the amount of pollutants emitted by CPPs in GAINS; E G is the total pollutant emissions from the power sector in GAINS.
Thus, we can determine the pollutant emission amount of the power sector after ULE technology is applied in CPPs: Similarly, we assumed that ULE technology can be used in ICB and achieve the same emission level as CPPs. Therefore, the emission can be obtained after the application of ULE technology in ICB, based on a similar calculation method. It is also assumed that the proportion of emissions from ICB in the industry sector on the MEIC is the same as the proportion of emissions from ICB in the industry sector on the GAINS, and the GAINS provides detailed data of pollutant classification emissions.
Finally, the new emission inventory is obtained after the application of the ULE technology in CPPs and ICB. The difference between the new emission and MEIC inventory is that it reveals the emission reduction of NO x , SO 2 and PM after applying ULE technology in CPPs and ICB. The emission reduction ratio of different pollutants after the application of ULE technology in CPPs and ICB can be obtained as follows: Here, R is the emission reduction ratio; E ULE is the total pollutant emissions of the new emission inventory.

Scenario Design
Seven simulation scenarios were designed by using the GEOS-Chem model to computationally estimate the environmental benefits of the application of ULE technology in CCPs and ICB, which are listed in Table 2. All scenarios used the spatial distribution and seasonal variation of the MEIC emission inventory, and we assumed the same meteorological field as the STD scenario to minimize the impact of weather effects. The period for these simulations was 16 December 2014 to 31 January 2015. December 2014 was used as the spin-up period to eliminate the influence of initial conditions. The simulations predicted the daily concentrations of PM 2.5 , sulfate, nitrate and ammonium. The average values for each of these components were computed using data from 1 January 2015 to 31 January 2015. January was selected as it is the most polluted month and can best reflect the environmental benefits brought about by ULE emission reduction. The standard scenario (STD) is based on the situations before implementing ULE facilities in January 2015. The remaining situations cover different scenarios of interest. The CPPs-ULE scenario assumes that all of China's CPPs have been retrofitted for ULE. The CPPs-ICB-ULE scenario assumes that the ICB process reaches the same emission level as that in the CPPs. The other four scenarios are used to further assess the environmental benefits of NH 3 emission reduction based on the CPPs-ULE scenario and CPPs-ICB-ULE scenario at a national level. NH 3 emission reduction refers to the total emission reduction of industries, residents, transportation and agriculture. The ammonia emission model (PKU-NH 3 ) used in the study has been proven to accurately describe the spatial distribution and seasonal characteristics of NH 3 emissions in China [18,25], and to simulate the changes in secondary inorganic aerosols in atmospheric chemical transmission models [26]. Regarding China's agricultural production, the excessive use of nitrogen fertilizer and extensive animal husbandry caused by NH 3 accounted for 80% of the country's total emissions, based on PKU-NH 3 model estimates, through the peak nitrogen fertilizer emissions during the growing season (spring and summer) to implement reasonable fertilization (i.e., avoid excessive fertilization and the implementation of deep fertilizer) and so on; optimizing agricultural management can achieve an NH 3 emission reduction of 50% [18]. In the field of livestock manure management, promoting the optimization of animal and poultry manure management, such as reducing the surface area and exposure time of the waste in the house, covered storage, low protein feeding and deep burial of feces at the depth of cultivated land, can significantly reduce NH 3 emissions [27,28]. If agriculture and animal husbandry take concerted measures to control NH 3 emissions, China's NH 3 emissions can reach more than 50%; therefore, this study selected NH 3 emission reductions of 30% and 50% to explore the potential environmental benefits.

Model Description and Sensitivity Tests
To assess the impacts of the emission reduction brought about by applying ULE facilities in CPPs and CIB on the ambient PM 2.5 concentration, several simulations using the GEOS-Chem (version 11-01) chemical transport model were conducted. To obtain more accurate results, the GEOS-Chem nested model at a high resolution of 0.25 • latitude × 0.3125 • longitude was used. Boundary conditions needed by the nested model were derived from the global atmospheric chemical model with a resolution of 2 • latitude × 2.5 • longitude every 3 h. Both the nested model and global model are driven by the GEOS-FP meteorology fields, which is from the National Aeronautics and Space Administration (NASA) Global Modeling Assimilation Office. Both models have 47 layers in the vertical direction, and the lowest 10 layers are almost 130 thick. Model convection is parameterized by the relaxed Arakawa-Schubert scheme [29] and the vertical mixing scheme in the planetary boundary layer employs a non-local scheme implemented by Lin and McElroy (2010) [30].
All simulations were run with the full chemistry mechanism (Ox-NOx-CO-VOC-HO x ) and aerosols, which include secondary inorganic aerosols (SIOA, including sulfate, nitrate and ammonium), black carbon (BC), primary organic carbon (POA), dust and sea salts, were calculated online. Modifications of the full chemistry mechanism followed Ni et al., (2018) [31]. The ammonium-sulfate-nitrate aerosol system, which uses the ISORROPIA II thermodynamic equilibrium model [32] to simulate SIOA, was coupled to the full chemistry mechanism and implemented in GEOS-Chem [33]. The parameterization of heterogeneous aerosol chemistry was treated as uptake coefficients [34]. Mineral dust aerosols were calculated online by the DEAD scheme [35] and the implementation of sea salt aerosol parameterization followed Jaegle et al., (2011) [36]. In addition to natural dust aerosols, anthropogenic sources of dust have been verified to play an important role in the total PM 2.5 concentration over China. Thus, the contribution of anthropogenic dust emissions was explicitly considered by adding the gridded monthly PM emission inventory from MEIC into GEOS-Chem.
Compared to observations, standard GEOS-Chem tends to underestimate the ambient concentration of sulfate due to the simplified sulfate production mechanism implemented in the model [37,38]. The standard GEOS-Chem model simulates sulfate production by considering the gas-phase oxidation of SO 2 by OH [39] and the aqueous-phase oxidation of S(IV) (including dissolvedSO 4 2− , HSO 3 − and SO 3 2− ) by hydrogen peroxide (H 2 O 2 ) and ozone (O 3 ) in cloud droplets. To comprehensively present the process of sulfate production,   [40] and   [41] assessed the heterogeneous uptake of SO 2 on the surface of deliquesced aerosols under high relative humidity conditions and the mechanism of the aqueous-phase oxidation of S(IV) by dissolved nitrogen dioxide (NO 2 ) in cloud droplets and aerosols with high relative humidity was followed. In addition, primary sulfate emissions are also important, accounting for 3.1% of the total Chinese anthropogenic sulfur emissions. The share of primary sulfate emissions over China could be underestimated; thus, the contribution was increased to 4.5%. These updates are welldocumented and are believed to bring the model predictions into closer agreement with measurements, especially for the winter months.
Global anthropogenic emissions were taken from EDGAR v4.2, with regional emissions overwritten by regional inventories such as the MIX Asian emission inventory over Asia [42], EMEP over Europe and NEI2011 over the US. Anthropogenic emissions over China of the standard scenario were taken from MEIC, except NH 3 emissions, which were from PKU-NH 3 [18]. Gridded monthly anthropogenic emissions for simulations conducted for other scenarios were customized and are summarized in Table 2. Parameterization of soil NO x emissions followed Hudman et al., (2012) [43]. Biomass burning emissions were taken from the Global Fire Emission Database version 4 (GFED v4) [44]. Biogenic non-methane volatile organic compound (NMVOC) emissions were calculated online by using the Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN v2.1) [45]. The model simulation results with hourly ground PM 2.5 concentration observation data were verified and evaluated, which concluded that the improved model used in this study can better calculate the concentration and spatial distribution of PM 2.5 in China on the whole, which also indicates that this model can be used for subsequent simulation studies [46].

Pollutant Reduction Potential
Before analyzing the environmental benefits of the application of ULE technology in CCPs and ICB by using GEOS-Chem, the pollutant reduction potential of ULE technology application in CPPs and ICB was analyzed. Based on MEIC, the total national NO x and SO 2 emissions in 2015 were calculated as 23.7 Tg and 18.2 Tg, respectively ( Figure 1). Among them, the amounts of NO x and SO 2 emitted by the power sector were 5.1 Tg and 4.0 Tg, representing 21.5% and 21.9%, respectively; the amounts of NO x and SO 2 emitted by the industrial sector were 9.7 Tg and 11.0 Tg, representing 40.9% and 60.4%, respectively. The ULE of CPPs in China (CHN) can reduce the NO x and SO 2 of CPPs themselves by 87.0% and 93.3%, while the emission reduction of the whole power plant sector is 83.3% and 91.6%, accounting for 17.9% and 20.0% of the total NO x and SO 2 emissions, respectively. If ULE facilities are installed both in CPPs and ICB, the total NO x and SO 2 can be reduced by 34.2% (8.1 Tg) and 51.0% (9.3 Tg), respectively. Due to the differences in the industrial structure of coal-based energy in various provinces and regions, under the CPPs-ULE and CPPs-ICB-ULE scenarios, the emission reductions of NO x and SO 2 are different. Environmental benefits are brought about by the reduction of NO x and SO 2 in regions such as Beijing-Tianjin-Hebei (BTH), the Yangtze River Delta (YRD), the Sichuan-Chongqing Economic Circle (SCC) and the Pearl River Delta (PRD), etc., from the promotion of ULE technology nationwide.

Regional Sector Emission Reduction
The reduction ratios of NO x and SO 2 in the power industry and the total of different regions under different scenarios are shown in Figure 2. The ICB-ULE refers to the realization of ULE transformation in the ICB process, which was only used to calculate the pollutant reduction and does not quantify the corresponding environmental benefits separately. As shown in Figure 2, under the CPPs-ULE scenario, the NO x and SO 2 emission reductions in the power sector in BTH, YRD, SCC and PRD reached 76.2%, 81.4%, 85.4%, 77.2% and 85.2%, 87.1%, 96.4%, 85.9%, respectively. It can be seen that the emissions of CPPs account for an absolute proportion of the emissions of the power sector, and the ULE of CPPs has greatly reduced the emissions of NO x and SO 2 from the power plants. The larger the reduction ratio of NO x and SO 2 in the power sector, the more urgent it is to promote the ULE transformation, and the more significant the environmental benefits will be. Under the ICB-ULE scenario, the NO x and SO 2 emission reductions of the industrial sectors in BTH, YRD, SCC and PRD reach 59.9%, 40.8%, 32.8%, 18.4% and 66.2%, 48.3%, 69.4%, 29.0%, respectively. The difference in the emission reduction ratio of NO x and SO 2 reflects the difference in industrial structure. For BTH, the emission reduction ratio is as high as 60%, indicating that the ICB emissions occupy a dominant position, while for PRD, the emission reduction ratio is less than 1/3, indicating that the ICB emissions occupy a relatively small proportion in the industrial structure. The emission reductions of NO x and SO 2 in the power and industrial sectors of BTH, YRD, SCC and PRD reach 64.3%, 55.0%, 42.6%, 42.1% and 68.6%, 58.3%, 73.5%, 45.9%, respectively, under the CPPs-ICB-ULE scenario.

Regional Emission Reduction
The reduction ratios of NO x and SO 2 in different regions under different scenarios are shown in Figure 3. The reduction ratios of NO x and SO 2 emissions in different regions under different scenarios.
As shown in Figure 3, under the CPPs-ULE scenario, the NO x and SO 2 emission reductions of the BTH, YRD, SCC and PRD regions reached 10.5%, 17.4%, 9.0%, 17.8% and 8.8%, 20.8%, 13.7%, 23.6%, respectively, in total NO x and SO 2 emissions in each region. It can be seen that the ULE of CPPs in the BTH and YRD region reduced the total NO x and SO 2 emissions in the region by around 10%, while the emission reduction of CPPs in the SCC and PRD region reduced the total NO x and SO 2 emissions by around 20%. Under the ICB-ULE scenario, the emission reductions of NO x and SO 2 in the industrial sectors of BTH, YRD, SCC and PRD reached 30.6%, 16.2%, 15.2%, 6.3% and 46.8%, 33.3%, 54.7%, 18.9%, respectively, in total NO x and SO 2 emissions in each region. Under the CPPs-ICB-ULE scenario, the emission reductions of NO x and SO 2 of BTH, YRD, SCC and PRD reached 41.1%, 33.6%, 24.2%, 24.0% and 55.6%, 54.1%, 68.4, 42.5%, respectively, in total NO x and SO 2 emissions in each region.
For the BTH region, the NO x and SO 2 emission reduction ratios for ICB-ULE are 2.9 and 5.3 times higher than CPPs-ULE, indicating that ICB-ULE in this region can significantly reduce pollutant emissions. For the YRD region, the NO x and SO 2 emission reduction ratios for ICB-ULE are 0.9 and 1.6 times those of the CPPs-ULE. For the SCC region, the NO x and SO 2 emission reduction ratios for ICB-ULE are 1.7 and 4.0 times higher than CPPs-ULE. For the PRD region, the NO x and SO 2 emission reduction ratios for ICB-ULE are 1.7 and 4.0 times higher than CPPs-ULE. For the PRD region, the NO x and SO 2 emission reduction ratios for ICB-ULE are 0.4 and 0.8 times those of the CPPs-ULE, indicating that the different emission reduction ratios of various regions are determined by differences in the energy industry structure.

Benefits of Primary Pollutant Emission Reduction
The spatial distribution of the concentrations of the primary pollutants, NO x and SO 2 , under different scenarios based on GEOS-Chem simulation is shown in Figure 4.  Figure 4 shows that under the STD scenario, the average NO x concentration in January 2015 in the North China Plain and the Yangtze River Delta was as high as 42-60 ppbv. Over other densely populated regions in Eastern China, the concentration of NO x was 6-24 ppbv, while, in the central and western regions of China, due to the relatively weak human activities, the emission of NO x was relatively low. This concentration was also very low, only 0.5-6 ppbv. The spatial distribution of the SO 2 concentration was slightly different from that of NO x . The difference lies in the high SO 2 concentration of 15-30 ppbv in Chongqing, Guizhou and Hunan province, which is caused by the energy structure. Under the CPPs-ULE scenario, the NO x and SO 2 emission reductions are mainly concentrated in high-value regions, and CPPs-ICB-ULE reduces the NO x and SO 2 emissions more significantly than CPPs-ULE. Figure 5 shows the changes in the average concentrations of NO x and SO 2 in different regions under different scenarios.
It can be seen from Figure 5 that, under the CPPs-ULE and CPPs-ICB-ULE scenarios, the concentration of NO x in the BTH region decreased from 24.9 ppbv to 20.7 ppbv and 13.5 ppbv, a decrease of 17.1% and 45.9%, respectively, while the concentration of SO 2 was reduced from 11.4 ppbv to 10.2 ppbv and 6.5 ppbv, a decrease of 10.7% and 42.7%, respectively, compared with the STD scenario. Similarly, the concentration of NO x in the YRD region was reduced from 19.9 ppbv to 14.9 ppbv and 11.9 ppbv, a decrease of 25.2% and 40.1%, and the concentration of SO 2 was decreased from 7.7 ppbv to 6.0 ppbv and 3.8 ppbv, a decrease of 22.0% and 51.0%, respectively. The concentration of NO x in the SCC region decreased from 4.6 ppbv to 3.8 ppbv and 3.3 ppbv, a decrease of 25.2% and 40.1%, and the concentration of SO 2 decreased from 8.9 ppbv to 7.3 ppbv and 4.8 ppbv, a decrease of 18.6% and 45.9%, respectively. The concentration of NO x in the PRD region was reduced from 6.8 ppbv to 5.3 ppbv and 4.6 ppbv, a reduction of 23.0% and 33.1%, and the concentration of SO 2 was reduced from 4.0 ppbv to 3.1 ppbv and 2.0 ppbv, a reduction of 23.0% and 49.4%, respectively. The concentration of NO x in the CHN region was reduced from 5.0 ppbv to 3.8 ppbv and 3.0 ppbv, a decrease of 24.7% and 39.8%, and the concentration of SO 2 was reduced from 3.6 ppbv to 2.9 ppbv and 2.2 ppbv, a decrease of 18.4% and 40.0%, respectively. By comparison with the actual emission reductions of NO x and SO 2 in the different regions in Figures 1 and 3, it can be seen that the emission reduction ratio of primary pollutants NO x and SO 2 is not consistent with the actual emission reduction ratio. For example, in the CPPs-ULE and CPPs-ICB-ULE scenarios, the actual NO x and SO 2 emission reduction ratios corresponding to BTH are 10.5%, 41.1%, 8.8% and 55.6%, while the corresponding primary pollutant emission reduction ratios are 17.1%, 45.9%, 10.7% and 42.7%, which is mainly because the actual NO x and SO 2 emitted into the atmosphere is also affected by atmospheric transport [47]. It is also affected by complex atmospheric chemical reactions [48,49].

Benefits of Secondary Pollutant Emission Reduction
The spatial distribution of the concentrations of the secondary pollutants, ammonium, nitrate, sulfate and PM 2.5 , under different scenarios based on GEOS-Chem simulation is shown in Figure 6.
It can be seen from Figure 6 that the high-value regions of ammonium, nitrate and sulfate in January 2015 are not consistent under the STD scenario. The high-value regions of ammonium are mainly concentrated in North China, Central China and the SCC regions. Among them, the high ammonium concentration in SCC is greater than 14 µg/m 3 , while in North and Central China, the high ammonium salt concentration is 6-14 µg/m 3 . The high concentration of nitrate is also mainly concentrated in North China, Central China and the SCC regions. The high-value region in North China has a concentration greater than 32 µg/m 3 , while that in the high-value region in SCC is relatively low, at 12-32 µg/m 3 . The concentration distribution of sulfate is mainly concentrated in the SCC region, and the concentration is usually 16-36 µg/m 3 . Ammonium, nitrate and sulfate are important components of PM 2.5 , so the high concentration of PM 2.5 is a combination of the three, mainly concentrated in the North China, Central China and SCC regions, where the PM 2.5 concentration in some regions is greater than 180 µg/m 3 . Under the CPPs-ICB-ULE scenario, the ICB has achieved ULE, further reducing NO x and SO 2 emissions, making the national emission reductions of ammonium, sulfate and PM 2.5 more obvious. The nitrate reduction regions are also concentrated in Northeast China, Henan and part of the western regions. Among them, the decrease in Henan is more obvious (reduced by 4.9%), while in Urumqi, North China, South China, SCC and PRD, the concentration of nitrate increased more significantly.
The formation of ammonium, nitrates and sulfates in the atmosphere is all related to NH 3 emissions, and there is a competition mechanism between nitrates and sulfates [17]: the ammonium ions (NH 4 + ) in the atmosphere are more likely to combine with sulfate ions (SO 4 2− ) to generate ammonium sulfate ((NH 4 ) 2 SO 4 ). Only when NH 3 is abundant will it continue to react with nitrate ions (NO 3 − ) to produce ammonium nitrate (NH 4 NO 3 ), so the reaction between inorganic substances has the following four situations: (a) the emission of NH 3  For regions with high NH 3 emissions, because NH 3 emissions are high, NH 4 + will consume all the SO 4 2− , but it is not enough to combine with all NO 3 − . When SO 2 is reduced, the NH 4 + originally combined with SO 4 2− will combine with the remaining NO 3 − to form NH 4 NO 3 , so the concentration of nitrate will increase, and there will be surplus after the combination of NH 4 + and NO 3 − , which cannot form ammonium. Thus, the concentration of ammonium will also decrease.
Because the sulfur content of coal used in different regions, power plant boiler technology, industrial coal use, agricultural NH 3 emissions, etc., are different, their individual NO x , SO 2 and NH 3 emission ratios are different, which makes the implementation of ULE reduction technology challenging. The emission reduction effects in the past were different, which is ultimately reflected in the different non-linear chemical mechanisms between ammonium, nitrate and sulfate in the atmosphere of various provinces and regions, which affects the environmental benefits brought about by emission reduction policies. Therefore, in order to better reduce the concentration of PM 2.5 , the SCC region should focus on the emission reduction of ammonium and sulfate precursors NH 3 and SO 2 , while, for the North and Central China areas, the focus should be on the reduction of ammonium and nitrate precursors NH 3 and NO x .
It can be seen from Figure 7 that, under the CPPs-ULE scenario, the emissions of primary pollutants NO x and SO 2 are significantly reduced (24.7% and 18.4%), causing the CHN concentrations of ammonium, nitrate, sulfate and PM 2.5 to decrease from 3.  In CPPs and ICB, the application of ULE reduction technology can greatly reduce the NO x and SO 2 emissions, and then decrease ammonium sulfate and PM 2.5 concentrations, but it increases the concentration of nitrate; therefore, the CHN regional NH 3 emissions are too high, and only NO x and SO 2 emissions are reduced; therefore, this is not the best method to reduce PM 2.5 measures, and other methods should be considered according to different regional condition NH 3 emissions.
Under the CPPs-ULE scenario, the BTH concentrations of ammonium, nitrate, sulfate and PM 2.5 decreased from 6.8 µg/m 3 , 14.0 µg/m 3 , 7.3 µg/m 3 and 75.8 µg/m 3 to 6.4 µg/m 3 , 14.0 µg/m 3 , 6.3 µg/m 3 and 73.4 µg/m 3 , respectively, which decreased by 5.0%, −0.2%, 13.2% and 3.1%, respectively. With the realization of ULE of CPPs in the BTH region, the emissions of primary pollutants NO x and SO 2 were significantly reduced (17.1% and 10.7%), leading to a significant decrease in the concentration of ammonium and sulfate. However, due to the competition mechanism between nitrate and sulfate, the concentration of nitrate was increased, but the proportion of the increase in nitrate concentration was relatively low (0.2%). Under the CPPs-ICB-ULE scenario, the emissions of primary pollutants NO x and SO 2 were reduced (45.9% and 42.7%), which could reduce the concentrations of ammonium, nitrate, sulfate and PM 2.5 in the BTH region to 5.7 µg/m 3 , 14.1 µg/m 3 , 4.3 µg/m 3 and 68.6 µg/m 3 , respectively, compared to the baseline scenario, which indicates reductions of 15.7%, −0.3%, 40.7% and 9.5%, respectively. It can be seen that for the BTH region, there are greater NH 3 emissions, but the region is not particularly affluent, because the concentration of nitrate under the different scenarios only increases by 0.2% and 0.3%.
Under the CPPs-ULE scenario, the emissions of primary pollutants NO x and SO 2 are significantly reduced (25.2% and 22.0%), causing the YRD concentrations of ammonium, nitrate, sulfate and PM 2.5 to decrease from 7.3 µg/m 3 , 14.0 µg/m 3 , 9.2 µg/m 3 and 72.8 µg/m 3 to 6.8 µg/m 3 , 14.5 µg/m 3 , 7.3 µg/m 3 and 69.5 µg/m 3 , respectively, which are decreased by 6.1%, −3.9%, 20.3% and 4.5%, respectively. Under the CPPs-ICB-ULE scenario, the emissions of primary pollutants NO x and SO 2 are reduced (40.1% and 51.0%), which can reduce the concentrations of ammonium, nitrate, sulfate and PM 2.5 in the YRD region to 6.2 µg/m 3 , 15.1 µg/m 3 , 5.0 µg/m 3 and 65.8 µg/m 3 , respectively, compared to the baseline scenario, with reductions of 14.7%, −8.2%, 45.9% and 9.6%, respectively. It can be seen that, compared with the BTH region, YRD is richer in NH 3 emissions. After the ULE of CPPs is completed, the nitrate concentration is increased by 3.9%. With the completion of the ULE of ICB, it has further reduced the emissions of NO X and SO 2 and has increased the concentration of nitrate to 8.2%. The YRD region pays more attention to NH 3 emission reduction control than the BTH region.
Under the CPPs-ULE scenario, the emissions of primary pollutants NO x and SO 2 are significantly reduced (15.2% and 18.6%), causing the SCC concentrations of ammonium, nitrate, sulfate and PM 2.5 to decrease from 7.9 µg/m 3 , 10.3 µg/m 3 , 13.5 µg/m 3 and 69.6 µg/m 3 to 7.2 µg/m 3 , 11.1 µg/m 3 , 10.8 µg/m 3 and 66.4 µg/m 3 , respectively, amounting to reductions of 8.8%, −7.5%, 20.0% and 4.6%, respectively. Under the CPPs-ICB-ULE scenario, the emissions of primary pollutants NO x and SO 2 are reduced (26.1% and 45.9%), which can reduce the concentrations of ammonium, nitrate, sulfate and PM 2.5 in the SCC region to 6.3 µg/m 3 , 11.7 µg/m 3 , 7.8 µg/m 3 and 62.4 µg/m 3 , respectively, compared to the baseline scenario, with reductions of 20.3%, −13.4%, 42.2% and 10.4%, respectively. It can be seen that, compared with the BTH and YRD regions, SCC has the most abundant NH 3 emissions. After the ULE of CPPs is completed, the nitrate concentration is increased by 7.5%. With the completion of the ULE of ICB, it has further reduced the emissions of NO X and SO 2 and has increased the concentration of nitrate to 13.4%. The SCC region is in urgent need of NH 3 emission control compared with the PRD region.
It can be seen that the PRD area has more NH 3 emissions than the BTH, but it is not richer than YRD and SCC. Therefore, after the ULE of CPPs is completed, the nitrate concentration is increased by 1.0%. With the completion of the ULE of ICB, it has further reduced the emissions of NO X and SO 2 and has increased the concentration of nitrate to 4.9%.

Environmental Benefits of NH 3 Emission Reduction
The spatial distribution of the concentrations of secondary pollutants, ammonium, nitrate, sulfate and PM 2.5 , under the NH 3 emission reduction control scenario based on GEOS-Chem simulation is shown in Figure 8.
As can be seen from Figure 8, under the CPPs-ULE-NH 3 -30% scenario, the ULE of CPPs with NH 3 emission reduced by 30% at the same time can cause the national ammonium, nitrate and PM 2.5 concentrations to significantly decrease, and the concentration of sulfate has no obvious change relative to the STD scenario. The main reason for this is the reduction of NH 3 in the atmosphere, which reduces the concentration of NH 4 + . Because of the non-linear mechanism of ammonium, nitrate and sulfate, NH 4 + is preferentially combined with SO 4 2− to form sulfate, and there is not enough NH 4 + to combine with NO 3 − to produce nitrate, so the concentration of ammonium and nitrate is significantly reduced. As the NH 3 emission reduction reaches 50%, the concentrations of ammonium, nitrate and PM 2.5 can still be further reduced, but this cannot cause a significant change in the concentration of sulfate. This shows that when NH 3 is reduced by 50%, the combination of NH 4 + and SO 4 2− is not affected. In order to better reduce PM 2.5 , the emission of one or more pollutants can be further reduced according to the cost of the emission reduction of NO x , SO 2 and NH 3 . On the basis of CPPs-ICB-ULE, the NH 3 emission reduction was 30% and 50%, respectively, making the changes of ammonium salt, nitrate, sulfate and PM 2.5 in the atmosphere consistent with the changes in the NH 3 emission reduction on the basis of CPPs-ULE. NH 3 emission reduction can significantly reduce the concentration of nitrate while reducing the ammonium and sulfate in PM 2.5 . As can be seen from Figure 8, under the CPPs-ULE-NH3-30% scenario, the ULE of CPPs with NH3 emission reduced by 30% at the same time can cause the national ammonium, nitrate and PM2.5 concentrations to significantly decrease, and the concentration of sulfate has no obvious change relative to the STD scenario. The main reason for this is the reduction of NH3 in the atmosphere, which reduces the concentration of NH4 + . Because of the non-linear mechanism of ammonium, nitrate and sulfate, NH4 + is preferentially combined with SO4 2− to form sulfate, and there is not enough NH4 + to combine with NO3 − to produce nitrate, so the concentration of ammonium and nitrate is significantly reduced. As the NH3 emission reduction reaches 50%, the concentrations of ammonium, nitrate and PM2.5 can still be further reduced, but this cannot cause a significant change in the concentration of sulfate. This shows that when NH3 is reduced by 50%, the combination of NH4 + and SO4 2− is not affected. In order to better reduce PM2.5, the emission of one or more pollutants can be further reduced according to the cost of the emission reduction of NOx, SO2 and NH3. On the basis of CPPs-ICB-ULE, the NH3 emission reduction was 30% and 50%, respectively, making the changes of ammonium salt, nitrate, sulfate and PM2.5 in the atmosphere consistent with the changes in the NH3 emission reduction on the basis of CPPs-ULE. NH3 emission reduction can significantly reduce the concentration of nitrate while reducing the ammonium and sulfate in PM2.5.
As can be seen from Figure 9, compared with the STD scenario, NH3 was reduced by 30% on the basis of CPPs-ULE, and the concentrations of CHN ammonium, nitrate, sulfate

Discussion
Although the CPPs have achieved the ULE transformation target, the PM 2.5 concentration across the country has decreased by 4.8% (1.4 µg/m 3 ) compared with January 2015, and haze pollution is still very serious. In addition, ammonium, nitrate and sulfate are important components of PM 2.5 . Due to the complex non-linear chemical competition mechanism among them, the average concentration of nitrate in CHN increases by 1.5% (0.1 µg/m 3 ) under the condition of a large reduction in NO x and SO 2 emissions, which reduces the environmental benefits of power plant emission reduction. If the ULE technology is applied to the ICB to further reduce NO x and SO 2 , although the PM 2.5 concentration can be reduced by 10.1% (2.9 µg/m 3 ), the concentration of nitrate will increase by 2.7% (0.2 µg/m 3 ). It can be seen that reducing the emission of NO x and SO 2 does not reduce the concentration of PM 2.5 well, and the haze pollution is still very serious.
On the basis of CPPs-ULE, NH 3 emissions are reduced by 30%, which can reduce the national PM 2.5 concentration by 11.5% (reduced by 3.3 µg/m 3 ), by significantly reducing the ammonium (reduced by 21.8%) and nitrate (reduced by 26.0%) concentrations, to achieve a better PM 2.5 reduction effect than ULE of ICB. If the NH 3 emission reduction reaches 50%, the concentrations of ammonium (reduced by 33.4%), nitrate (reduced by 46.8%) and sulfate can be further reduced, and the PM 2.5 concentration can be reduced by 16.5% (4.7 µg/m 3 ). Similarly, on the basis of CPPs-ICB-ULE, NH 3 emissions are reduced by 30% and 50%, which can reduce the national PM 2.5 concentration by 15.6% (4.4 µg/m 3 ) and 20.3% (5.8 µg/m 3 ), by significantly reducing the ammonium and nitrate concentrations.
In conclusion, NO x , SO 2 and NH 3 emission reduction have the same PM 2.5 reduction effect. ULE of CPPs and ICB reduces the PM 2.5 concentration by reducing NO x and SO 2 , and then reducing the concentrations of ammonium and sulfate, while NH 3 emission reduction mainly reduces the PM 2.5 concentration by reducing the concentration of ammonium and nitrate. In order to better reduce the concentration of PM 2.5 , the emission reduction control measures of NO x , SO 2 and NH 3 should be formulated according to the specific conditions of different regions.
As can be seen from Figure 10, for the BTH region in the CPPs-ULE scenario, as with the NO x and SO 2 emission reduction, the concentrations of ammonium and sulfate were reduced by 5.0% and 13.2% respectively, while the concentration of nitrate increased by 0.2%, and the concentration of PM 2.5 was reduced by 3.1% (2.5 µg/m 3 ). Under the CPPs-ICB-ULE scenario, with the further decrease in NO x and SO 2 , the concentrations of ammonium and sulfate decreased by 15.7% and 40.7%, respectively, while the concentration of nitrate increased by 0.3%, and the concentration of PM 2.5 decreased by 9.5% (7.2 µg/m 3 ). Therefore, relative to the CPPs-ULE scenario, reducing the emissions of NH 3 can significantly reduce the concentrations of nitrate and ammonium, and reduces the sulfate concentration by 1.6%. This indicates that the atmospheric NH 3 emission in the BTH region is insufficient, and NH 4 + is insufficient to combine with all SO 4 2− . When the NH 3 emission is reduced to 50%, the concentrations of ammonium, nitrate and sulfate can be reduced by 33.0%, 42.8% and 16.5%, respectively, and the concentration of PM 2.5 can be reduced by 13.9% (10.6 µg/m 3 ). On the basis of CPPs-ICB-ULE, the emission of NH 3 was reduced by 30%, the concentrations of ammonium, nitrate and sulfate were reduced by 27.5%, 18.4% and 41.5%, respectively, and the concentration of PM 2.5 was decreased by 14.1% (10.7 µg/m 3 ). With the reduction of NH 3 to 50%, the concentrations of ammonium, nitrate and sulfate were decreased by 38.6%, 36.0% and 42.5%, respectively, and the concentration of PM 2.5 was decreased by 18.4% (14.0 µg/m 3 ). In conclusion, after the completion of ULE of CPPs in the BTH region, the synergistic effect of NH 3 emission reduction should be comprehensively considered while promoting ULE technology in the field of ICB. For the YRD region in the CPPs-ULE scenario, as with the NO x and SO 2 emission reduction, the concentrations of ammonium and sulfate were reduced by 6.1% and 20.3%, respectively, while the concentration of nitrate increased by 3.9% (0.5 µg/m 3 ), and the concentration of PM 2.5 was reduced by 4.5% (3.3 µg/m 3 ). Under the CPPs-ICB-ULE scenario, with the further decrease in NO x and SO 2 , the concentration of ammonium and sulfate decreased by 14.7% and 45.9%, respectively, while the concentration of nitrate increased by 8.2% (1.1 µg/m 3 ), and the concentration of PM 2.5 decreased by 9.6% (7.0 µg/m 3 ). Therefore, relative to the CPPs-ULE scenario, an emission reduction of NH 3 by 30% can significantly reduce the concentrations of nitrate and ammonium, and also reduces the sulfate concentration by 0.7%. This indicates that the atmospheric NH 3 emission in the YRD region is insufficient, and NH 4 + is insufficient to combine with all SO 4 2− . When the NH 3 emission is reduced to 50%, the concentrations of ammonium, nitrate and sulfate can be reduced by 39.0%, 49.5% and 21.7%, respectively, and the concentration of PM 2.5 can be reduced by 18.2% (13.3 µg/m 3 ). On the basis of CPPs-ICB-ULE, the emission of NH 3 was reduced by 30%, the concentrations of ammonium, nitrate and sulfate were reduced by 30.6%, 19.5% and 46.3%, respectively, and the concentration of PM 2.5 was decreased by 16.5% (12.0 µg/m 3 ). With the reduction of NH 3 to 50%, the concentrations of ammonium, nitrate and sulfate were decreased by 43.7%, 41.7% and 46.7%, respectively, and the concentration of PM 2.5 was decreased by 22.1% (16.1 µg/m 3 ). In conclusion, relative to the BTH region, the emission of NH 3 in the YRD region is higher, severely weakening the reduction effect of ULE in CPPs and ICB. After the completion of ULE of CPPs, an NH 3 emission reduction by 30% can achieve a better PM 2.5 reduction effect than that of ULE of ICB. Therefore, after the completion of ULE of CPPs, priority should be given to the environmental benefits brought about by NH 3 emission reduction, and then to ULE of ICB.
For the SCC region in the CPPs-ULE scenario, as with the NO x and SO 2 emission reduction, the concentrations of ammonium and sulfate were reduced by 8.8% and 20.0%, respectively, while the concentration of nitrate increased by 7.5% (0.8 µg/m 3 ), and the concentration of PM 2.5 was reduced by 4.6% (3.2 µg/m 3 ). Under the CPPs-ICB-ULE scenario, with the further decrease in NO x and SO 2 , the concentration of ammonium and sulfate decreased by 20.3% and 42.2%, respectively, while the concentration of nitrate increased by 13.4% (1.4 µg/m 3 ), and the concentration of PM 2.5 decreased by 10.4% (7.2 µg/m 3 ). Therefore, relative to the CPPs-ULE scenario, an emission reduction of NH 3 by 30% can significantly reduce the concentrations of nitrate and ammonium, and also reduces the sulfate concentration by 0.3%. This indicates that the atmospheric NH 3 emission in the SCC region is insufficient, and NH 4 + is insufficient to combine with all SO 4 2− . When the NH 3 emission is reduced to 50%, the concentrations of ammonium, nitrate and sulfate can be reduced by 37.2%, 59.1% and 20.4%, respectively, and the concentration of PM 2.5 can be reduced by 17.8% (12.4 µg/m 3 ), relative to the STD scenario. On the basis of CPPs-ICB-ULE, the emission of NH 3 was reduced by 30%, the concentrations of ammonium, nitrate and sulfate were reduced by 32.4%, 17.4% and 42.4%, respectively, and the concentration of PM 2.5 was decreased by 16.3% (11.4 µg/m 3 ). With the reduction of NH 3 to 50%, the concentrations of ammonium, nitrate and sulfate were decreased by 43.3%, 45.0% and 42.5%, respectively, and the concentration of PM 2.5 was decreased by 21.7% (15.1 µg/m 3 ).
In conclusion, the relative to the YRD region, the emission of NH 3 in SCC region is higher, severely weakening the reduction effect of ULE in CPPs and ICB. After the completion of ULE of CPPs, an NH 3 emission reduction by 30% can achieve a better PM 2.5 reduction effect than that of ULE of ICB, and an NH 3 emission reduction by 50% can achieve a better PM 2.5 reduction effect than that of ULE of ICB with NH 3 emissions reduced by 30%. Therefore, after the completion of ULE of CPPs, priority should be given to the environmental benefits brought about by NH 3 emission reduction, and then to ULE of ICB.
For the PRD region in the CPPs-ULE scenario, as with the NO x and SO 2 emission reduction, the concentrations of ammonium and sulfate were reduced by 11.4% and 18.8%, respectively, while the concentration of nitrate increased by 1.0% (0.1 µg/m 3 ), and the concentration of PM 2.5 was reduced by 6.2% (2.7 µg/m 3 ). Under the CPPs-ICB-ULE scenario, with the further decrease in NO x and SO 2 , the concentrations of ammonium and sulfate decreased by 23.9% and 40.5%, respectively, while the concentration of nitrate increased by 4.9% (0.3 µg/m 3 ), and the concentration of PM 2.5 decreased by 12.7% (5.6 µg/m 3 ). Therefore, relative to the CPPs-ULE scenario, an emission reduction of NH 3 by 30% can significantly reduce the concentrations of nitrate and ammonium, and also reduces the sulfate concentration by 0.2%. This indicates that the atmospheric NH 3 emission in the PRD region is insufficient after a 30% reduction in NH 3 , and NH 4 + is insufficient to combine with all SO 4 2− . When the NH 3 emission is reduced to 50%, the concentrations of ammonium, nitrate and sulfate can be reduced by 35.3%, 59.7% and 19.2%, respectively, and the concentration of PM 2.5 can be reduced by 17.0% (7.5 µg/m 3 ), relative to the STD scenario. On the basis of CPPs-ICB-ULE, the emission of NH 3 was reduced by 30%, the concentrations of ammonium, nitrate and sulfate were reduced by 36.5%, 29.5% and 40.6%, respectively, and the concentration of PM 2.5 was decreased by 18.6% (8.2 µg/m 3 ). With the reduction of NH 3 to 50%, the concentrations of ammonium, nitrate and sulfate were decreased by 45.3%, 53.1% and 40.6%, respectively, and the concentration of PM 2.5 was decreased by 22.7% (10.1 µg/m 3 ).
In conclusion, the emission of NH 3 in the PRD region is relatively high, severely weakening the reduction effect of the ULE in CPPs and ICB. After the completion of ULE of CPPs, an NH 3 emission reduction by 30% can achieve the same PM 2.5 reduction effect as that of ULE of ICB, and an NH 3 emission reduction by 50% is weaker than that of ULE of ICB with NH 3 emissions reduced by 30% for a PM 2.5 reduction effect. Therefore, after the completion of ULE of CPPs, the formulation of environmental policies should refer to the BTH region, and the synergistic effect of NH 3 emission reduction should be comprehensively considered during the ULE of ICB. Based on the competitive mechanism of sulfate and nitrate, the emission reduction of NH 3 makes the NH 3 content in the atmosphere change from "surplus" to "insufficient", and the concentration of nitrate is significantly reduced in the process of transformation to ammonia deficiency, followed by the concentrations of ammonium and sulfate, which is consistent with the conclusions of other scholars [50]. NH 3 emission reduction converts NH 4 SO 4 into NH 4 HSO 4 . In addition, SO 2 reacts in various ways to form sulfate [48], so, with the increase in the NH 3 emission reduction ratio, the sulfate concentration decreases slowly. The main reason for the decrease in the nitrate and ammonium concentrations caused by NH 3 emission reduction is the reduction in NH 3 available for reaction in the atmosphere and the reduction in precursors of this reaction. Other related chemical reaction mechanisms need to be further studied [49,51,52].

Conclusions
Although the CPPs have achieved the ULE transformation target, the PM 2.5 concentration across the country has decreased by 4.8% (1.4 µg/m 3 ) compared with January 2015, and the haze pollution is still very serious. In addition, ammonium, nitrate and sulfate are important components of PM 2.5 . Due to the complex non-linear chemical competition mechanism among them, the average concentration of nitrate in the country increases by 1.5% (0.1 µg/m 3 ) under the condition of a large reduction in NO x and SO 2 emissions, which reduces the environmental benefits of ULE of CPPs. If the ULE technology is applied to the ICB to further reduce NO x and SO 2 , the PM 2.5 concentration can be reduced by 10.1% (2.9 µg/m 3 ), and the concentration of nitrate will increase by 2.7% (0.2 µg/m 3 ). Therefore, it is concluded that simply reducing NO x and SO 2 emissions cannot sufficiently reduce the concentration of PM 2.5 . Targeted pollutant emission reduction control measures should be formulated according to the specific conditions of different regions to achieve a better reduction in PM 2.5 concentration.
Based on the CPPs-ULE, an NH 3 emission reduction of 30% can significantly reduce the concentrations of ammonium and nitrate, so that the PM 2.5 concentration decreases by 11.5% (3.3 µg/m 3 ), to achieve a better PM 2.5 reduction effect than ULE of ICB. If the NH 3 emission reduction reaches 50%, the concentrations of ammonium, nitrate and sulfate can be further reduced, and the PM 2.5 concentration can be reduced by 16.5% (4.7 µg/m 3 ). Similarly, based on the CPPs-ICB-ULE, NH 3 is reduced by 30% and 50%, and the PM 2.5 concentration is reduced by 15.6% (4.4 µg/m 3 ) and 20.3% (5.8 µg/m 3 ). In summary, NO x , SO 2 and NH 3 emission reduction have the same PM 2.5 reduction effect. The ULE of CPPs and ICB can reduce NO x and SO 2 , thereby reducing the concentrations of ammonium and sulfate, causing the PM 2.5 concentration to be reduced, and NH 3 reduction is achieved mainly through reducing the concentrations of ammonium and nitrate to reduce the concentration of PM 2.5 . In order to better reduce the concentration of PM 2.5 , NO x , SO 2 and NH 3 emission reduction control measures should be comprehensively considered in different regions of China.
The emission bases of NO x , SO 2 and NH 3 in different regions of China are different, so the changes in regional PM 2.5 and its important components under different scenarios are also quite different. For the BTH and PRD regions, after the ULE of CPPs is completed, the synergistic effects of NH 3 emission reduction should be comprehensively considered while promoting ULE reduction technologies in the field of ICB. For the YRD and SCC regions, after the completion of ULE of CPPs, priority should be given to the environmental benefits brought about by NH 3 emission reduction, and then to the ULE transformation of ICB. In conclusion, in order to reduce PM 2.5 more effectively, the economic cost and environmental benefits of emission reduction control measures for NO x , SO 2 , NH 3 and other pollutants should be taken into account comprehensively.