Predicted impact of thermal power generation emission control measures in the Beijing-Tianjin-Hebei region on air pollution over Beijing, China

Widespread economic growth in China has led to increasing episodes of severe air pollution, especially in major urban areas. Thermal power plants represent a particularly important class of emissions. Here we present an evaluation of the predicted effectiveness of a series of recently proposed thermal power plant emission controls in the Beijing-Tianjin-Hebei (BTH) region on air quality over Beijing using the Community Multiscale Air Quality(CMAQ) atmospheric chemical transport model to predict CO, SO2, NO2, PM2.5, and PM10 levels. A baseline simulation of the hypothetical removal of all thermal power plants in the BTH region is predicted to lead to 38%, 23%, 23%, 24%, and 24% reductions in current annual mean levels of CO, SO2, NO2, PM2.5, and PM10 in Beijing, respectively. Similar percentage reductions are predicted in the major cities in the BTH region. Simulations of the air quality impact of six proposed thermal power plant emission reduction strategies over the BTH region provide an estimate of the potential improvement in air quality in the Beijing metropolitan area, as a function of the time of year.

SCIEnTIfIC REPORTS | (2018) 8:934 | DOI: 10.1038/s41598-018-19481-0 resulting from targeted SO 2 and NO x emission controls for 2010; it was predicted that annual PM 2.5 and SO 2 could decline by 3-15 µg m −3 (4-25%) and 30-60%, respectively. Wang et al. 16 predicted that for areas with PM 2.5 concentrations exceeding the second-level air quality standard (25 µg m −3 for annual mean and 50 µg m −3 for 24-hour mean) in 31 selected provinces in China, the annual mean SO 2 and NO x concentrations in 2020 relative to 2010 could be reduced by 40.0% and 31.6%, respectively, and the annual mean PM 2.5 concentration could decrease by 17.2%. Here we present results of the application of the CMAQ air quality model (see Methods) to evaluate the effects of thermal power generation emission controls in the Beijing-Tianjin-Hebei (BTH) region ( Fig. 1). Table 1 lists the emission controls corresponding to a range of policies for thermal power plants in the BTH region. All thermal power plants in the BTH region use coal as the only fuel. These include the policy A "Emission standard of air pollutants of thermal power plants (GB13223-2011)" (released on July 29, 2011) and policy B "Action Plan for the Transformation and Upgrading of Coal Power Energy Conservation and Emission Reduction (2014-2020)" (released by the Chinese government on September 12, 2014) (http://www.sdpc.gov.cn/ gzdt/201409/t20140919_626240.html). The eight emission scenarios in Table 1 will be evaluated here for thermal power plants in the BTH region. Since the policy A was carried out on January 1, 2013, it is assumed that the baseline scenario, Case 1, was implemented as of that date. Case 1 therefore serves as a basis: (1) to simulate current air quality in the BTH region for the base year; (2) to evaluate model performance; and (3) to determine the effectiveness of other potential policies to improve air quality in Beijing. According to policy A, key areas such as the BTH region need to conform to special emission limits. Key areas are defined as those with high land development density, weakened atmospheric carrying capacity, and fragile ecological environment. In Case 2 for  special emission limits, the emission standards for SO 2 , smoke dust, and PM 2.5 as listed in Table 1 are changed to 50, 20, and 10 mg m −3 , respectively. According to the policy B, all thermal power units must meet ultra-low emission standards. In Case 3, emission standards of SO 2 , NO 2 , smoke dust, and PM 2.5 are 70%, 50%, 50%, and 80% of those of Case 2, respectively. Policy B includes acceleration of the upgrade of active thermal power units, especially for those <200 MW. Figure 2 shows the distribution and percentage of the thermal power plants in terms of unit capacities in the three areas for the BTH region in 2013 (http://ieimodel.org/jjjdqhdhypfqd). Units with capacity <200 MW and >600 MW account for 64% and 7% of the total units, respectively (Fig. 2B), while the unit capacity is mainly located in Tianjin, Shijiazhuang, Tangshan, and Handan, especially in the southwest and east of the BTH region ( Fig. 3A and B). In Case 4, all units with output <200 MW are assumed to be eliminated to evaluate the extent to which this structural adjustment would affect Beijing air quality.

Emission Control Scenarios
The Shenhua Group Corporation Ltd. proposed three emission control policies including "Extremely ultra-low" (Case 5), "Near zero" (Case 6), and "Green power generation" (Case 7) (http://www.shenhuagroup. com.cn/shjtww/1382682123426/201506/cc7d7362c2224378ab536a27aa30b8b7.shtml). The Shenhua Group is the largest coal supplier in the world. In Case 5, the emission standards for smoke dust and PM 2.5 are set to be one-half, but remain unchanged for SO 2 and NO 2 , relative to Case 3. In Case 6, emission standards of SO 2 , NO 2 , smoke dust, and PM 2.5 are set at 40%, 50%, 20%, and 20% of those of Case 5, respectively. In Case 7,emission standards for SO 2 , NO 2 , smoke dust, and PM 2.5 lie between those of Cases 5 and 6. Case 8 represents the hypothetical total absence of thermal power plants in the BTH region. Since air quality in Beijing is poorest in the winter, simulations for Cases 5, 6, and 7 are carried out only for the month of January over the domain d02 with the grid resolution of 12 km (Fig. 1).
Predicted influence of emission control policies on Beijing air quality. Monthly mean concentrations of PM 2.5 , PM 10 , NO 2 , CO, and SO 2 for different emission scenarios at 12 monitoring stations in the Beijing are used as a basis to assess the predicted improvement in Beijing air quality associated with the emission control  Table 2 summarizes the predicted annual percentage reductions in Cases 2, 3, 4 relative to Case 1 (CO, SO 2 , NO 2 , PM 2.5 , and PM 10 ), while Table 3 shows the percentage reductions for Cases 5, 6, 7 in January, and Table 4 shows the percent reductions for total removal of power plants. Spatial distributions of the reduction of PM 2.5 for January, April, July, and October are shown in Fig. 4, while Fig. 5 shows the spatial distributions of predicted monthly mean reductions for PM 2.5 for implementation of Cases 5, 6, 7 relative to that of Case 1. Cases 2 and 3 lead to higher reduction percentages than Case 4 for all species except CO, for which all three emission control policies predict similar reduction percentages (Table 2). Predicted monthly mean reductions in Beijing in January for Cases 5, 6, 7 for PM 2.5 range from 9.05% to 12.12% (Table 3). The spatial distributions of predicted reduction of PM 2.5 in different months for Cases 2, 3, 4, shown in Fig. 4, indicate that the largest percent reduction of PM 2.5 would occur in January, especially in Tianjin and over the central part of Hebei (southwest of Beijing), for Cases 2 and 3. This response is consistent with the dominant locations of the coal-fired power plants (Fig. 3A). To meet increasingly stringent emission standards and obtain maximum emission reductions from the coal-fired power plants, "Extremely ultra-low" (Case 5), "Near zero" (Case 6), and "Green power generation" (Case 7) policies have been proposed. Emission levels from coal-fired power plants adopting these newly-designed emission control technologies are comparable with those from natural gas-fired plants. Table 3 summarizes predicted reduction percentages and amounts of PM 2.5 , PM 10 , NO 2 , CO, and SO 2 in Beijing in January for these cases. Spatial distributions of PM 2.5 reduction amounts over the BTH region in Fig. 5 show that most of PM 2.5 reduction amounts are located in the southeast Beijing, Tianjin and the central Hebei with the largest reductions from Cases 6 ("Near zero emission") and 8 ("No thermal power plants") and lowest reductions from Case 5 ("Extremely ultra-low"), being consistent with the emission control standards (Table 1) and domain locations of the coal-fired power plants (Fig. 3A), as expected.
Predicted SO 2 reduction percentages for the newly-designed emission control policies (Cases 5, 6, 7) are predicted to range from −9.2% to −14.8%, exceeding those of the current emission control policies (ranging from −2.7% to −6.0%). As expected, among these emission policies, Case 6 ("Near zero emission") leads to the largest reduction percentages for all species (PM 2.5 , PM 10 , NO 2 , CO, and SO 2 ) (Table 3), consistent with the emission control standards in Table 1. For Case 8 "No thermal power plants" in Table 4, the predicted annual contributions of thermal power plants over the BTH region to the concentrations of CO, SO 2 , NO 2 , PM 2.5 , and PM 10 in Beijing are 37.6%, 23.1%, 23.0%, 23.8% and 24.0%, respectively. Thermal power plants over the BTH region are predicted to contribute appreciably to CO concentrations in Beijing ( Table 4). The lowest values in Beijing contributed by thermal power plants over the BTH region are predicted to occur in April for CO, SO 2 , and PM 2.5 , whereas in July for NO 2 and PM 10 .
In comparison with the prediction of the current emission control policies (Cases 2, 3 and 4) in Table S2, Cases 5, 6 and 7 lead to further modest reductions of PM 2.5 , PM 10 , NO 2 , CO and SO 2 in Beijing (see Table S3). For example, the reduction percentages for PM 2.5 for the newly-designed emission control policies lie between −9.1% and −12.1%, about 10% to 7% higher than those of the current emission control policies (range from −2.0% to −5.1%) in January Predictions for PM 10 are close to those for PM 2.5 . CO reduction percentages for the newly-designed emission control policies range from −23.0% and −27.7%, about 3% to 7% higher than those of the current emission control policies (range from −20.4% to −20.8%), and for SO 2 , reduction percentages for the newly-designed emission control policies range from −9.2% and −14.8%, about 6% to 12% higher than those of the current emission control policies (range from −2.7% to −6.0%). As expected, among these newly-designed emission policies, Case 6 ("Near zero emission") predicts the largest reduction percentages for all species (PM 2.5 , PM 10 , NO 2 ,CO and SO 2 ), followed by Cases 7 and 5 (see Table S3), consistent with the emission control standards in Table 1. The results in Tables S2 and S3 suggest that the extent to which it is worth carrying out these newly-designed emission control policies depends on their economics.
Tables S10-S21 (Supplemental Information) summarize predicted annual PM 2.5 , CO, SO 2 , NO 2 , and PM 10 reduction percentages in 12 other major Chinese cities (Tianjin, Baoding, Cangzhou, Chengde, Handan, Hengshui, Qinhuangdao, Shijiazhuang, Tangshan, Xingtai, Zhangjiakou, Langfang) over the BTH region for Cases 2, 3, 4 on the basis of simulations at the 12 km grid resolution. The annual mean predicted reduction percentages for PM 2.5 are between −4.5% and −10.7%, with the highest value in Zhangjiakou and the lowest value in Hengshui for Cases 2 and 3, while they lie between −1.7% and −3.4% for Case 4. Predicted reduction percentages for Case 3 are slightly higher than those of Case 2 for all species (CO, SO 2 , NO 2 , PM 2.5 , and PM 10 ) for all cities except Tianjin. Predicted monthly mean reduction amounts in January are the highest for CO, SO 2 , NO 2 , PM 2.5 , and PM 10 for all three emission control policies, as expected.

Conclusions
Thermal power plants (entirely coal-burning) are a major source of atmospheric emissions in China, controls on which are crucial for the improvement of air quality. In this work, we assess the potential air quality benefits in Beijing from different thermal power emission control policies for the BTH region. Predicted annual mean reduction percentages in Beijing lie between −5.3% and −6.3% for PM 2.5 , PM 10 , NO 2 , and SO 2 for Cases 2 ("Special emissions limits") and 3 ("Ultra-low emission standards"), and between −2.2% and −3.0% for Case 4 ("Adjustment of structures"), reflecting the effects of eliminating all units <200 MW. All three emission control   [18][19][20][21][22] . We use a nested grid configuration with an outer grid encompassing most of China and part of eastern Asia (36 km grid resolution), the first inner grid encompassing the North China Plain (12 km grid resolution) and the second inner grid covering the BTH region (4 km grid resolution) (Fig. 1). The physics package of the WRF3.4 (ARW) includes version 2 of the Kain-Fritsch cumulus cloud parameterization (KF2) 23 , Morrison et al. two-moment cloud microphysics [24][25][26] , the Asymmetric Convective Model version 2 (ACM2) for the planetary boundary layer (PBL) 27,28 , the RRTMG radiation mechanism and the Pleim-Xiu land-surface model 29,30 , with indirect soil moisture and temperature nudging 31,32 . The aerosol module of the CMAQ model is AERO6, and the gas-phase chemical mechanism is CB05. Boundary conditions for the inner domains are derived from simulations of the outer domains and the meteorological initial, and lateral boundary conditions for the outermost domain were derived from the National Center for Environmental Prediction (NCEP) final analysis dataset with a spatial resolution of 1° × 1° and a temporal resolution of 6 h. The default chemical boundary conditions (BCONs) in the CMAQ model were used in the simulations for the outermost domain of base year 2013. The WRF-CMAQ model simulation periods include January, April, July, and October in 2013 to represent winter, spring, summer, and autumn seasons, respectively.
Model performance for PM 2.5 , PM 10 , NO 2 , SO 2 , and CO. To evaluate model performance, the mean bias (MB), normalized MB (NMB) and root mean square error (RSME), normalized mean error (NME) and correlation coefficient (r) are calculated 33 . Monthly and annual results of model performance evaluation for PM 2.5 , PM 10 , NO 2 , CO, and SO 2 in Beijing are summarized in Table S1 (see Supplemental Information) for the baseline emission scenario (Case 1) at the grid resolution of 12 km. Figure 6 shows the time-series comparisons of the observed and predicted hourly mean PM 2.5 concentrations in Beijing for each month at the grid resolution of 12 km. Model performance for PM 2.5 , PM 10 , NO 2 , CO, and SO 2 in Beijing is similar for the simulations at grid resolutions of 36, 12, and 4 km (see Supplemental Information). For example, the NMB values for PM 2.5 are 19.6%, 26.6%, and 23.2% at 36 km, 12 km, and 4 km grid resolutions, respectively, on the basis of the annual simulations, while the corresponding NMB values for PM 10 are 3.6%, 10.9%, and 8.1%, respectively.
Simulations at all three grid resolutions for SO 2 , CO, and PM 2.5 exhibit poorer performance for July and October as compared with that for January and April (see Supplemental Information). Model simulations at all three grid resolutions (4 × 4, 12 × 12, 36 × 36 km) exhibit good performance for NO 2 for all months except July. Significant overestimation of SO 2 in July is likely a result of non-representative locations and elevations of surface  34 , where SO 2 is primarily emitted from stacks above local shallow inversion layers, while measurement stations are located close to the surface. Time-series comparisons of observed and predicted PM 2.5 for different months in Beijing (Fig. 6) indicate that the predictions capture the hourly variations and broad synoptic changes in the observed PM 2.5 concentrations.
Hourly observed concentrations (PM 2.5 , PM 10 , NO 2 , CO, and SO 2 ) at 12 monitoring stations in Beijing obtained from the website "China's air quality on-line monitoring analysis platform (http://www.aqistudy.cn/)" were used for evaluating the two-way coupled WRF-CMAQ model. The 00°N, 116.41°E). Since observational data before 18:00 January 17, 2013 are unavailable, the period from January 18 to February 17 of 2013 was used to represent January 2013. To evaluate model performance, concurrent hourly predicted concentrations at the monitoring sites were averaged in parallel with the hourly observations.