The Aerosol-Radiation Interaction E ﬀ ects of Di ﬀ erent Particulate Matter Components during Heavy Pollution Periods in China

: The Beijing-Tianjin-Hebei (BTH) region experienced heavy air pollution in December 2015, which provided a good opportunity to explore the aerosol-radiation interaction (ARI) e ﬀ ects of di ﬀ erent particulate matter (PM) components (sulfate, nitrate, and black carbon (BC)). In this study, ﬁve tests were conducted by the Weather Research and Forecasting—Chemistry (WRF-Chem) model. The tests included scenario 1 simulation with ARI turned on, scenario 2 simulation with ARI turned o ﬀ , scenario3 simulation without NO x / NO 3 − emissions and with ARI turned on, scenario 4 simulation without SO 2 / SO 42 − emissions and with ARI turned on, and scenario 5 simulation without BC emissions and with ARI turned on. The ARI decreased the downward shortwave radiation (SWDOWN) and the temperature at 2 m (T2), reduced the planetary boundary layer (PBL) height (PBLH), and increased the relative humidity (RH) at 2 m in the region. These factors also contribute to pollution accumulation. The results revealed that BC aerosols have a stronger e ﬀ ect on the reduction in SWDOWN than sulfate (SO 42 − ) and nitrate (NO 3 − ). BC aerosols produce both cooling and heating e ﬀ ects, while SO 42 − aerosols produce only cooling e ﬀ ects. The PBL decreased and RH2 increased due to the aerosol feedback e ﬀ ect of sulfate, nitrate, and BC. The ARI e ﬀ ect on meteorological factors during the nonheavy pollution period was much smaller than that during the pollution period.


Introduction
Aerosols can affect the temperature and relative humidity (RH) at the surface by scattering and absorbing solar radiation in the atmosphere [1][2][3], which is called the aerosol-radiation interaction (ARI). The ARI effect plays an important role in both regional meteorological variations and climate change. It can alter the vertical mixing of mass and momentum in the planetary boundary layer (PBL) and perturb meteorological variables such as surface temperature, wind, and planetary boundary layer height (PBLH). Moreover, ARIs also change the photolysis rates of photochemical and regional meteorological factors, which affect air pollutant dispersion [4][5][6][7][8][9]. The effect of the ARI is different among different aerosol components. Sulfate aerosols suspended in the atmosphere appear in the form of sulfuric acid (SO 4 2− ), which has a cooling effect and decreases the temperature at the surface. Black carbon (BC) has both cooling and heating effects and is emitted into the atmosphere by the combustion of solid fuels (e.g., wood, crop residues, and coal), biomass fuels, and fossil fuels. Anthropogenic emissions worldwide contribute to aerosol pollutants (e.g., sulfate, nitrate, and organic carbon). Together, they produce a cooling effect with a total decreased direct radiative forcing of 0.5 W m −2 and indirect cloud albedo forcing of 0.7 W m −2 at the surface [10][11][12]. Air pollution has a great impact and is a considerable threat to human health, and it also affects the climate. In China, particulate matter (PM) pollution is a severe environmental problem [13]. The Chinese government has made great efforts to improve air quality and reduce air pollution and has also implemented a series of air pollution control measures, which have effectively controlled regional PM 2.5 pollution and improved air quality. For example, the Ministry of Environmental Protection of China released the "Air Pollution Prevention and Management Plan for the Beijing-Tianjin-Hebei region and its Surrounding Areas 2017" [14]. The plan suggests that the Beijing, Tianjin, and Hebei Provinces, as well as surrounding provinces (i.e., Shanxi Province, Shandong Province, and Henan Province), will constitute a regional network to reduce extremely high PM 2.5 concentrations. When heavy air pollution is expected, emission reduction measures will be carried out simultaneously in these areas to prevent air pollution. To ensure good air quality during major events, such as the 70th anniversary of the Anti-Japanese War victory and the Asia-Pacific Economic Cooperation, emission controls have been implemented in these regions (the Beijing, Tianjin, Hebei, and Shaanxi provinces). Those measures were implemented successfully, and the pollution levels were low [15,16]. These practices provided opportunities to experiment and explore the atmospheric chemical mechanism in China. For example, the primary pollutants were significantly reduced, but the secondary pollutants increased after the full emission controls during the 2008 Olympic Games [17,18]. Secondary aerosols had the greatest reductions, while the primary aerosols experienced a smaller change during the Asia-Pacific economic (APEC) period [19,20]. However, only a few studies have focused on the ARI effect in China during those emission control implementation periods.
Previous studies have paid attention to the potential feedback of aerosols to meteorological factors. Makar et al. [21] found that temporal and spatial variations in meteorological elements were due to direct and indirect aerosol feedbacks, with the largest effect occurring in the summer near large pollution emission sources. Forkel et al. [22] found that changes in PBLH and temperature cannot always be related to the distribution of aerosol concentration, since changes in clouds dominate the direct effect of aerosol particles on solar radiation. The regional model found that the ARI effect and aerosol-cloud interaction with anthropogenic SO 4 2− induces a negative radiative forcing, which results in cooling temperatures at the surface and a decrease in precipitation over East Asia [23]. The regional climate chemistry model system investigated single scattering albedo forcing and BC loading on the climate [24]. The study on aerosol feedbacks in Beijing, China, showed increases in recent years. Yang et al. [25] found that the maximum reduction in wind speed by aerosols in Beijing is approximately 3.1%, corresponding to a change in RH-corrected visibility of approximately 10 km. During a haze episode in December 2012 in Beijing, Quan et al. [26] observed that PBLH decreased from 1.5 km to approximately 0.5 km in the most serious period of haze. Gao et al. [27] simulated that aerosols lead to a negative radiative forcing of −20 to −140 W m −2 at the surface, temperature decreases by 0.8-2.8 • C, and RH increases by approximately 4-12% at the surface in Beijing and Tianjin. Zhou et al. [28] found that the ARI can affect the attribution of PM 2.5 variability to emission changes and meteorological conditions in Beijing. However, there are no studies that have investigated the ARI effects of different PM components during heavy pollution periods over the most polluted region (Beijing-Tianjin-Hebei (BTH)) in China. Our knowledge of the ARI effect in China is far from complete. Knowing how these interactions may affect air pollution and meteorological factors will help to reduce extremely high Atmosphere 2020, 11, 254 3 of 18 PM 2.5 pollution. The feedback induced by different components must be studied. This will help to understand the complex relationship between air quality and meteorological factors. In December 2015, the northern region of China, especially the BTH area, experienced several episodes of heavy PM 2.5 pollution. These heavy pollution scenarios provided experimental opportunities to explore how different PM components (nitrate aerosols, sulfate aerosols, and BC) respond to ARI effects. In this paper, a fully coupled online model, the Weather Research and Forecasting-Chemistry (WRF-Chem) model, is used to carry out studies during weeks with pollution in China. This work presents a regional-scale sensitivity study performed by the WRF-Chem model, and several air pollution concentrations and meteorological parameters are analyzed. The research results of this paper provide a scientific basis for ARI studies and support for the Chinese government to prevent air pollution.

Model Description
In this study, the WRF-Chem model (version 3.4.1: National oceanic and atmospheric administration, United States) was used to study the ARI effect. The chemistry carbon bond mechanism version Z (CBM-Z) [29], coupled with the Model for Simulating Aerosol Interactions and Chemistry (MOSAIC) [30], was used. The CBM consists of 67 prognostic species and 164 reactions, which are used to calculate the gas-phase chemistry. The MOSAIC aerosol module includes methane, sulfate, sulfonate, chloride, nitrate, ammonium, sodium, BC, primary organic mass, liquid water, and other inorganic masses. The particle size distributions are divided into four size bins (0.039-0.1 µm, 0.1-1.0 µm, 1.0-2.5 µm, and 2.5-10 µm). The model calculates biogenic emissions online using the Gunther scheme. The physics options include the new Thompson microphysics option [31], the Goddard shortwave option [32], the Rapid Radiative Transfer Model (RRTM) longwave radiation option [33], and the Yonsei University (YSU) PBL option [34]. The ARI effect on shortwave radiation was based on the Mie theory, which follows the approach of Fast et al. [35]. The configuration is shown in Table 1.

Simulation Configurations and Design
The modeling domain covered a portion of northern China with 223 × 202 horizontal grid cells (as shown in Figure 1). The horizontal resolution was 9 km. The modeling vertical resolution was divided into thirty logarithmic structure layers, which ranged from the surface to the layer with a pressure of 100 mb. The National Centers for Environmental Prediction (NCEP) final reanalysis data were used to generate the initial meteorological conditions and boundary files.
In December, the BTH region experienced two episodes of heavy PM 2.5 pollution. The first episode was from 8 December to 10 December. The second episode was from 19 December to 22 December. When the mean hourly PM 2.5 concentrations were over 150 µg m −3 and lasted over three days, emergency emission reduction measures were implemented immediately in Beijing and its surrounding regions. These measures included suspending all construction projects, implementing even and odd-numbered license plate policies, and suspending the operation of more industrial plants. The 0.25 • Multi-resolution Emission Inventory for China (MEIC) emission inventory of 2014 was used as the base anthropogenic emission input [36]. In the two red alert periods, the emission reduction inventory was mainly updated based on the implementation of emergency control measures by municipal environmental protection bureaus. The simulation started on 25 November and ended on 31 December 2015. The first five days were excluded from the analysis and were considered as the spin-up time. In this research, five sensitivity simulations were run to investigate the ARI effect of different PM components (nitrate aerosol, sulfate aerosol, and BC). Table 2 summarizes the different scenarios.  [36]. In the two red alert periods, the emission reduction inventory was mainly updated based on the implementation of emergency control measures by municipal environmental protection bureaus. The simulation started on 25 November and ended on 31 December 2015. The first five days were excluded from the analysis and were considered as the spin-up time. In this research, five sensitivity simulations were run to investigate the ARI effect of different PM components (nitrate aerosol, sulfate aerosol, and BC). Table 2 summarizes the different scenarios.

Model Configuration Scenario 1
Real emission scenario; ARI turned on Scenario 2 Real emission scenario; ARI turned off Scenario 3 No NO3 − and NOx emissions; ARI turned on Scenario 4 No SO4 2− and SO2 emissions; ARI turned on Scenario 5 No BC emission reduction; ARI turned on The direct aerosol feedback in the scenario1 was activated and designed to represent the actual pollution process of which the simulation results were also used for model verification. The scenario 2 simulation had the same emissions as the scenario1, but the ARI option was turned off. The scenario3 simulation had no NO3 − or NOx emission reductions, and the ARI option was turned on. The scenario4 simulation had no SO4 2− or SO2 emission reductions, and the ARI option was turned on. The scenario5 simulation had no BC emissions, and the ARI option was turned on. The following are definitions of the ARI effects of nitrate aerosols, sulfate aerosols, and BC:  The direct aerosol feedback in the scenario1 was activated and designed to represent the actual pollution process of which the simulation results were also used for model verification. The scenario 2 simulation had the same emissions as the scenario1, but the ARI option was turned off. The scenario3 simulation had no NO 3 − or NO x emission reductions, and the ARI option was turned on. The scenario4 simulation had no SO 4 2− or SO 2 emission reductions, and the ARI option was turned on. The scenario5 simulation had no BC emissions, and the ARI option was turned on. The following are definitions of the ARI effects of nitrate aerosols, sulfate aerosols, and BC: Atmosphere 2020, 11, 254 5 of 18 where ∆V represents the effects of ARI; ∆BC EF represents the ARI effects of BC; ∆SO EF represents the ARI effects of sulfate aerosols; ∆NO EF represents the ARI effects of nitrate aerosols; and V represents the surface downward shortwave radiation (SWDOWN), temperature at 2 m (T2), relative humidity at 2 m (RH2), or PM 2.5 concentration.

Model Performance
In this study, meteorological factors were obtained from the China Meteorological Administration observation network, namely, the Meteorological Information Comprehensive Analysis and Process System (MICAPS). The monitoring data were collected from five MICAPS sites to evaluate the meteorological simulation performances (Beijing site: lat. 116. 28 . This website publishes hourly air quality information for 367 monitored cities in China. The hourly air pollutant concentrations from the CNEMC were collected to evaluate the simulation performances. The model was validated by comparing the results of the scenario1 simulation and the observed results for surface PM 2.5 concentrations, T2, 10 m wind speed (WS10), and RH2. The normalized mean gross error (NME), normalized mean bias (NMB), and correlation coefficient (RC) were used for the statistical analysis based on a previous study [37] and the U.S. EPA model evaluation protocol [38].
The evaluation statistics of T2 (K), RH2 (%), and WS10 (m s −1 ) for Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan are summarized in Table 3. Figure 2 presents the hourly simulated and observed meteorological variables and the PM 2.5 results. The WRF-Chem simulation results adequately captured the variations in T2 in these regions, with correlation coefficients of 0.72, 0.75, 0.71, 0.53, and 0.78 in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively. The NMB and NME indicated good model performance for T2. The NMB values of T2 were between −1.00% and 1.00%, and the NME values of T2 were 1.00%, 0.58%, 0.81%, 0.98%, and 0.75% in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively. As shown in Table 3, Beijing, Tianjin, and Tangshan showed the smallest NMB and NME values for the RH simulation. Baoding and Shijiazhuang presented relatively large biases for the RH simulation, and they are located in heavily polluted areas. The RH2 simulation results had correlation coefficients between 0.59 and 0.74. The simulation results adequately captured the variations in WS10. All correlation coefficients were greater than 0.5. However, as the analysis nudging option was not applied in the model, overpredictions occurred for hourly WS10 in the regions, with average bias values between 17.9 and 67.44 at five sites. The evaluation of the PM 2.5 concentrations in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan is also shown in Table 3. The NMB values for the comparison results of PM 2.5 were −41.37%, 24.66%, 28.35%, 30.95%, and −34.79% in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively. The NME values of PM 2.5 from the different sites were generally between 32.83% and 43.76%. The correlation coefficients of the PM 2.5 concentrations between the simulated and observed values were 0.67, 0.58, 0.80, 0.74, and 0.80 in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively. The model simulation results can clearly represent the air pollution process. The high uncertainty of emission inventories and meteorological simulation results affected the accuracy of the air pollution simulation results. A previous study also showed a large bias during heavily polluted days, which is an inherent characteristic that the model can produce [39]. According to the U.S. Environmental Protection Agency [38], all parameters followed the guidelines, and the WRF-Chem model predicted variables reasonably well in this work.    shows that PM 2.5 concentrations are negatively correlated with wind speed, which contributes to the dispersion of PM 2.5 pollutants. PM 2.5 concentrations are positively correlated with RH, which means the precursors (SO 2 , NO x , NH 3 , and volatile organic compounds (VOCs)) have the tendency to convert to PM 2.5 through chemical reactions. Figure 4 shows the spatial distributions of the PM 2.5 emission inventory, the mean simulation, and the monitor concentration in the BTH region over the simulation period. The cities of Shijiazhuang, Baoding, and Beijing and the southern part of the BTH region had high pollution emissions and heavy air pollution. Heavy pollution formed in the southern parts of the region (e.g., Shijiazhuang, Xingtai, and Handan) and spread towards the northern parts of the region (e.g., Zhangjiakou and Chengde). Eventually, pollution accumulated in front of the Taihang and Yanshan Mountains in the northern and western parts of the region.

ARI Effects of Different PM Components
The scenario1 and scenario2 model sensitivity results were compared to examine the ARI effects on meteorological variables (SWDOWN, T2, RH2, and PBLH).    The scenario1 and scenario2 model sensitivity results were compared to examine the ARI effects on meteorological variables (SWDOWN, T2, RH2, and PBLH). The ARI effects caused by NO3 − , SO4 2− and BC were also examined by comparing the results of scenario1 with the results of scenario3, the results of scenario1 with the results of scenario4, and the results of scenario1 with the results of scenario5. Based on the PM2.5 standard (75 µg m −3 ) in China, the ARI effects of PM components under different PM2.5 pollution levels were calculated. Two different PM2.5 pollution levels were examined (i.e., daily PM2.5 concentrations from 0 to 75 µg m −3 and greater than 75 µg m −3 ).

Downward Shortwave Radiation
The ARI effects on SWDOWN were examined, whose spatial distributions over the region in December are shown in Figure 5. The left-hand graphs in Figure 5 are the mean distributions of the meteorological variables from the scenario1 simulation. The right-hand graphs in the figure are the relative differences between the scenario1 and scenario2 simulations. The spatial mean contributions of the ARI effect to SWDOWN average were decreases of 14.83 W m −2 , 14.48 W m −2 , 18.53 W m −2 , 16.19 W m −2 , and 12.26 W m −2 in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, over the simulation period. The ARI decreased the SWDOWN in the region, particularly in areas with severe PM2.5 pollution. The SWDOWN decreased in five cities due to the ARI effects of BC, SO4 2− , and NO3 − , as shown in Figure 6. The BC absorption effect on solar radiation was stronger than the effects of sulfate and nitrate. In December, BC reduced shortwave radiation, of which averages of 7

Downward Shortwave Radiation
The ARI effects on SWDOWN were examined, whose spatial distributions over the region in December are shown in Figure 5. The left-hand graphs in Figure 5 are the mean distributions of the meteorological variables from the scenario1 simulation. The right-hand graphs in the figure are the relative differences between the scenario1 and scenario2 simulations. The spatial mean contributions of the ARI effect to SWDOWN average were decreases of 14.83 W m −2 , 14.48 W m −2 , 18.53 W m −2 , 16.19 W m −2 , and 12.26 W m −2 in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, over the simulation period. The ARI decreased the SWDOWN in the region, particularly in areas with severe PM 2.5 pollution. The SWDOWN decreased in five cities due to the ARI effects of BC, SO 4 2− , and NO 3 − , as shown in Figure 6. The BC absorption effect on solar radiation was stronger than the effects of sulfate and nitrate.       (scenario1) (scenario1-scenario2)/scenario1*100%

Temperatures at 2 m
As aerosols reduce incoming solar radiation by scattering and absorption, the surface temperatures decrease. T2 was reduced by up to 1.5 • C in the southern BTH region, of which the cooling effect occurred over most parts of the region, as shown in Figure 7 and Tangshan, respectively, due to nitrate. The temperatures fell by averages of 0.19 • C, 0.20 • C, 0.20 • C, 0.12 • C, and 0.07 • C in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, due to sulfate. BC produces both cooling and heating effects, which was reported by a previous study [12]. On the one hand, when a BC aerosol enters cloud droplet, it enhances the radiation absorption ability of the cloud droplet and leads to a temperature increase. On the other hand, it also promotes the cloud-forming process by acting as cloud condensation nodules and improves cloud reflection. Thus, BC aerosols can cause a surface cooling effect. In this paper, the temperatures fell by an average of 0.13 • C, 0.06 • C, 0.15 • C, 0.09 • C, and 0.00 • C in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, which was due to BC. The monthly mean ARI effects on T2 under different PM 2.5 pollution levels in the five cities are provided in Table 5. During the pollution period, the temperatures fell by averages of 0.94 • C, 0.25 • C, 0.5 • C, and 0.14 • C due to the TARI, SO 4 2− , NO 3 − , and BC, respectively, in the five cities. The temperatures fell by averages of 0.28 • C, 0.06 • C, 0.13 • C, and 0.04 • C due to the TARI, SO 4 2− , NO 3 − , and BC, respectively, in the five cities during the nonheavy pollution period.
The warming effect of BC aerosols caused the temperature to increase by 0.04 • C in Tangshan when the PM 2.5 concentration was higher than 75 µg m −3 .
temperatures decrease. T2 was reduced by up to 1.5 °C in the southern BTH region, of which the cooling effect occurred over most parts of the region, as shown in Figure 7. T2 was reduced by 0.69 °C, 0.63 °C, 0.72 °C, 0.55 °C, and 0.44 °C in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, over the simulation period, as shown in Figure 7. T2 decreased in the five cities due to the aerosol feedback effect caused by NO3 − , SO4 2− , and BC, as shown in Figure 8. The temperatures fell by averages of 0.36 °C, 0.36 °C, 0.33 °C, 0.24 °C, and 0.24 °C in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, due to nitrate. The temperatures fell by averages of 0.19 °C, 0.20 °C, 0.20 °C, 0.12 °C, and 0.07 °C in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, due to sulfate. BC produces both cooling and heating effects, which was reported by a previous study [12]. On the one hand, when a BC aerosol enters cloud droplet, it enhances the radiation absorption ability of the cloud droplet and leads to a temperature increase. On the other hand, it also promotes the cloud-forming process by acting as cloud condensation nodules and improves cloud reflection. Thus, BC aerosols can cause a surface cooling effect. In this paper, the temperatures fell by an average of 0.13 °C, 0.06 °C, 0.15 °C, 0.09 °C, and 0.00 °C in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively, which was due to BC. The monthly mean ARI effects on T2 under different PM2.5 pollution levels in the five cities are provided in Table 5. During the pollution period, the temperatures fell by averages of 0.94 °C, 0.25 °C, 0.5 °C, and 0.14 °C due to the TARI, SO4 2− , NO3 − , and BC, respectively, in the five cities. The temperatures fell by averages of 0.28 °C, 0.06 °C, 0.13 °C, and 0.04 °C due to the TARI, SO4 2− , NO3 − , and BC, respectively, in the five cities during the nonheavy pollution period. The warming effect of BC aerosols caused the temperature to increase by 0.04°C in Tangshan when the PM2.5 concentration was higher than 75 µg m −3 .

Plant Boundary Layer Height
The ARI reduced the PBLH in the region by 5~30 m, as shown in Figure 9. The mean contribution of the ARI to the PBLH in December saw between a 6.56% to 10   (scenario1) (scenario1-scenario2)/scenario1*100%

2-m Relative Humidity
The monthly average RH2 exhibited obvious increases in the middle and southern parts of the BTH region, as shown in Figure 11. In Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, the mean spatial contributions of the ARI to RH2 over the simulation period were approximately 1.93%, 3.61%, 3.67%, 2.81%, and 2.22%, respectively. The RH2 increased in the five cities due to the aerosol feedback effects of NO 3 − , SO 4 2− , and BC, as shown in Figure 12. Nitrate increased RH2 by averaging 0.92%, 1.50%, 1.42%, 1.07%, and 0.81% in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively. Sulfate increased RH2 by averages of 0.40%, 0.68%, 0.80%, 0.47%, and 0.39% in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively. BC increased RH2 by averages of 0.23%, 0.93%, 1.03%, 0.16%, and 0.28% in Beijing, Tianjin, Baoding, Shijiazhuang, and Tangshan, respectively. The effect of the TARI increased the RH2 in the region by an average of 4.40% during the pollution period and increased that by an average of 1.24% during the nonheavy pollution period (shown in Table 7). Sulfate aerosols increased the average RH2 by an average of 0.80% during the pollution period and 0.07% during the nonpollution period. Nitrate and BC aerosols increased RH2 by averages of 1.69% and 1.05%, respectively, during the pollution period, and 0.53% and 0.05%, respectively, during the nonpollution period.
respectively. The effect of the TARI increased the RH2 in the region by an average of 4.40% during the pollution period and increased that by an average of 1.24% during the nonheavy pollution period (shown in Table 7). Sulfate aerosols increased the average RH2 by an average of 0.80% during the pollution period and 0.07% during the nonpollution period. Nitrate and BC aerosols increased RH2 by averages of 1.69% and 1.05%, respectively, during the pollution period, and 0.53% and 0.05%, respectively, during the nonpollution period.
(scenario1) (scenario1-scenario2)/scenario1*100% Figure 11. Simulated results of the ARI effect on RH2.  The results in this paper are similar to those of previous studies. For example, in East China, the monthly mean SWDOWN, T2, and PBLH can decrease up to −12.37 W m −2 , −0.24°C, and −31.59 m due respectively. The effect of the TARI increased the RH2 in the region by an average of 4.40% during the pollution period and increased that by an average of 1.24% during the nonheavy pollution period (shown in Table 7). Sulfate aerosols increased the average RH2 by an average of 0.80% during the pollution period and 0.07% during the nonpollution period. Nitrate and BC aerosols increased RH2 by averages of 1.69% and 1.05%, respectively, during the pollution period, and 0.53% and 0.05%, respectively, during the nonpollution period.   The results in this paper are similar to those of previous studies. For example, in East China, the monthly mean SWDOWN, T2, and PBLH can decrease up to −12.37 W m −2 , −0.24°C, and −31.59 m due  The results in this paper are similar to those of previous studies. For example, in East China, the monthly mean SWDOWN, T2, and PBLH can decrease up to −12.37 W m −2 , −0.24 • C, and −31.59 m due to ARIs [40]. Even in the eastern continental United States, the ARI effect decreased the SWDOWN by 11.3 W m −2 , the T2 by 0.16 K, and the PBLH by 22.4 m in winter [41].

PM 2.5 Concentrations
The ARI reduced the surface temperature, increased RH2, and decreased PBLH, which led to a more stable lower atmosphere. Stable meteorological conditions suppressed the dispersion of air