The Development of an Atmospheric Aerosol/Chemistry‐Climate Model, BCC_AGCM_CUACE2.0, and Simulated Effective Radiative Forcing of Nitrate Aerosols

This study developed a next‐generation atmospheric aerosol/chemistry‐climate model, the BCC_AGCM_CUACE2.0. Then, the performance of the model for nitrate was evaluated, and the nitrate direct radiative forcing (DRF) and effective radiative forcing (ERF) due to aerosol‐radiation interactions were simulated for the present day (2010), near‐term future (2030), and middle‐term future (2050) under the Representative Concentration Pathway 4.5, 6.0, and 8.5 scenarios relative to the preindustrial era (1850). The model reproduced the distributions and seasonal changes in nitrate loading well, and simulated surface concentrations matched observations in Europe, North America, and China. Current global mean annual loading of nitrates was predicted to increase by 1.50 mg m−2 relative to 1850, with the largest increases occurring in East Asia (9.44 mg m−2), Europe (4.36 mg m−2), and South Asia (3.09 mg m−2). The current global mean annual ERF of nitrates was −0.28 W m−2 relative to 1850. Due to global reductions in pollutant emissions, the nitrate ERF values were predicted to decrease to −0.17, −0.20, and −0.24 W m−2 in 2030 and −0.07, −0.18, and −0.19 W m−2 in 2050 for Representative Concentration Pathway 4.5, 6.0, and 8.5 relative to 1850, respectively. Although global mean nitrate values showed a declining trend, future nitrate loading remained high in East Asia and South Asia.


Introduction
Anthropogenic emissions of aerosols and their precursors into the atmosphere have had a significant effect on the global environment and climate (Bian et al., 2017;Dong et al., 2010). As scattering aerosols, nitrate aerosols are effective at scattering solar radiation, and nitrate gaseous precursors are emitted in large quantities (Gao et al., 2004;Myhre et al., 2006;Zhang, Shen, et al., 2012). Several studies have shown that NO x (NO 2 and NO) and NH 3 emissions, the main precursors of nitrates, have continued to increase, resulting in a greater proportion of the total anthropogenic aerosols being composed of nitrates (Adams et al., 2001;Bellouin et al., 2011;Bian et al., 2017;Boucher et al., 2013;Hauglustaine et al., 2014;Liao & Seinfeld, 2005;Li et al., 2015;Myhre et al., 2006). Observations have shown that concentrations of nitrate in the atmosphere exceed those of sulfate in some regions of Europe and are higher in very industrialized areas (Malm et al., 2004;Putaud et al., 2004). Therefore, it is important to understand changes in nitrate concentrations and their contribution to radiative forcing in the future. Some studies have used global models to simulate present-day (PD) nitrate DRF values; however, predictions range widely (−0.025 to −0.19 W m −2 ) with high uncertainty (Andreae, 1995;Adams et al., 2001;Jacobson, 2001;Bauer et al., 2007;Bellouin et al., 2011;Hauglustaine et al., 2014;Liao et al., 2004;Liao & Seinfeld, 2005;Li et al., 2015;Myhre et al., 2013;Xu & Penner, 2012). The Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR5) gave an estimate of nitrate DRF values (−0.11 W m −2 [−0.3 to −0.03]) (Boucher et al., 2013), and other studies have predicted nitrate DRF values in the future. Under the Special Report on Emissions Scenarios A2, Adams et al. (2001) and Liao and Seinfeld (2005) predicted nitrate DRF values during 2100 of −1.28 and −0.95 W m −2 , using the Goddard Institute for Space Studies general circulation model GISS GCMIIprime and GISS GCM II′, respectively. Bauer et al. (2007) investigated nitrate DRF values in 2030 using the GISS model following the Special Report on Emissions Scenarios A1B and predicted that the global mean annual DRF value will be −0.09 W m −2 relative to the preindustrial era (PI). Hauglustaine et al. (2014) calculated DRF values due to nitrates of −0.06 to −0.11 W m −2 using the LMDz-INCA global model according to the different RCP scenarios. The global mean annual nitrate DRF values in 2090 were predicted to be −0.024, −0.038, and −0.047 W m −2 for the RCP4.5,6.0,and 8.5 scenarios, respectively (Li et al., 2015).
Although many studies have investigated the DRF values of nitrate, few have considered the optical properties of nitrates while using an atmospheric chemistry-climate online model (Adams et al., 2001;Andreae, 1995;Liao et al., 2004 ;van Dorland et al., 1997 ;Wang et al., 2010). Previous studies of nitrate DRF were susceptible to errors (Shen, 2009;Zhang, Shen, et al., 2012) as a result of using the optical properties of sulfate (Liao et al., 2004;van Dorland et al., 1997;Wang et al., 2010). IPCC AR5 first proposed the concept of effective radiative forcing (ERF), and few studies on nitrate ERFari (effective radiative forcing due to aerosolradiation interactions) are current. In this study, a next-generation atmospheric aerosol/chemistry-climate model, the BCC_AGCM_CUACE2.0 (the Atmospheric General Circulation Model of Beijing Climate Center Version 2 coupled with the China Meteorological Administration Unified Atmospheric Chemistry Environment Version 2) was developed, and calculated nitrate optical properties were applied within it. After evaluating the performance of the model for nitrates, the model was used to investigate the temporal and spatial variation in nitrate aerosols and DRF and ERF values from 1850 to 2010, 2030, and 2050 under three RCP scenarios, RCP4.5,6.0,and 8.5. As we didn't consider the aerosol-cloud interactions, the ERF refers to ERFari in this work.

Model Description 2.1.1. Atmospheric General Circulation Model
We used a next-generation atmospheric aerosol/chemistry-climate model BCC_AGCM_CUACE2.0 (the Atmospheric General Circulation Model of Beijing Climate Center Version 2 coupled with the China Meteorological Administration Unified Atmospheric Chemistry Environment Version 2). BCC_AGCM2.0 was developed by the National Climate Center of the China Meteorological Administration (NCC/CMA; Wu et al., 2010). BCC_AGCM2.0 has been participated in the IPCC Coupled Model Intercomparison Project Phase 5 and has been validated to be state-of-the-art against satellite observations (e. g., Jiang et al., 2015, Zhang, Shen, et al., 2012. The model has a horizontal resolution of T42 (approximately 2.8°× 2.8°) and 26 vertical levels, with a rigid lid at about 2.9 hPa. A more detailed dynamics, physical processes, and validation of BCC_AGCM2.0 are described in Wu et al., 2010. The cloud overlap scheme of the Monte Carlo independent column approximation (McICA) (Jing & Zhang, 2012Pincus et al., 2003) and the new Beijing Climate Center radiation transfer model BCC_RAD (Zhang, 2016;Zhang et al., 2014) were implemented in the model to improve its simulations of radiation fluxes at the top of the atmosphere (TOA) and the surface (Zhang et al., 2018). In BCC_RAD, the optical properties of aerosols were divided into 17 spectral bands from 0.204 to 1,000 μm (Zhang, 2016).

Unified Atmospheric Chemistry Environment
CUACE is a unified atmospheric chemistry environment, which includes the aerosol module (CUACE/ Aero), the gas chemistry module (CUACE/Gas), and the thermodynamic equilibrium module (CUACE/ ISO). CUACE/Aero, a size-segregated multicomponent aerosol module, was developed by Gong et al. (2002Gong et al. ( , 2003. The aerosol processes include emission, transport, chemical transportation, interactions with cloud, and deposition of aerosols. The model can calculate the mass concentration of the seven aerosol species, that is, sea salt, sand/dust, BC, OC, sulfate, nitrate, and ammonium salt (Zhou et al., 2012). Sulfates, BC, OC, sand/dust, nitrates, and sea salts are segregated into 12 size bins with radii between 0. 005-0.01, 0.01-0.02, 0.02-0.04, 0.04-0.08, 0.08-0.16, 0.16-0.32, 0.32-0.64, 0.64-1.28, 1.28-2.56, 2.56-5.12, 5.12-10.24, and 10.24-20.48 μm. The hygroscopic growth of soluble aerosol particles, such as nitrate, sulfate, OC, and sea salt, was taken into account. The aerodynamic size of the particles was taken from Kohler theory to be in equilibrium with the ambient relative humidity . When calculating the optical properties of one specie, we didn't take into account its mixing with other aerosols; however, for total optical property of all the aerosols, we will assume they are externally mixed.
CUACE/Gas is based on the second generation of Reginal Acid Deposition Model (RADM2) (Stockwell et al., 1990). The model consists of 63 gaseous species through 21 photochemical reactions and 121 gas phase reactions and is applicable under a wide variety of environmental conditions, especially for smog. CUACE/Gas provides the production rates of sulfate and second organic aerosols (SOAs) and generates an oxidation background for aqueous-phase chemistry of aerosols, in which sulfate transformation changes the distribution of SO 2 in clouds (Zhou et al., 2012). More details about specific parameterization schemes and algorithms are showed by Stockwell et al. (1990) and Zhou et al. (2012).
The nitrate and ammonium formed through the gaseous oxidation are unstable and prone to being decomposed back to their precursors. Thus, CUACE adopts ISORROPIA to calculate the thermodynamic equilibrium between them and their gas precursors (West et al., 1998;Nenes et al., 1998, Yu et al., 2005, known as CUACE/ISO in our model. A more detailed description of ISORROPIA can be found in Nenes et al., 1998.

Aerosol Optical Properties
Mie scattering theory was used to obtain the aerosol optical properties, that is, extinction cross section, single scattering albedo, and asymmetry factor. In this work, nitrate refractive index was from the HITRAN 2004 database (Rothman et al., 2005), and the uncertainty of aerosols refractive indices in the HITRAN database is less than 5% (e.g., Lund Myhre et al., 2005). The refractive index of dry nitrate particles at 0.55 μm was 1.398−10 −8 i, and the corresponding values for other aerosols were shown in . Based on the Mie scattering code of Wiscombe (1980), optical properties of nitrates can be calculated.
Nitrate, sulfate, OC, and sea salt particles are hygroscopic, and their optical properties will change with relative humidity. In this study, the relative humidity was first divided into 10 bins: 0, 45%, 50%, 60%, 70%, 80%, 90%, 95%, 98%, and 99%. Then the radii of wet aerosol particles in each size bin are obtained for these relative humidity bins. The wet particle densities and refractive indices consisting of real and imaginary parts are obtained according to volume-weighted method, respectively (Zhang, Wang, Niu, et al., 2012;. New size-segregated aerosol optical properties were calculated for all 17 spectral bands and each of the 10 relative humidity bins using Mie scattering theory (Wiscombe, 1980). These were then incorporated into the BCC_AGCM2.0. As an example, Figure 1 shows the changes in extinction cross section, single scattering albedo, and asymmetry factor of nitrate with relative humidity at λ = 0.55 μm when the dry radius of nitrate particle equals to 0.06 μm. The aerosol optical properties for any relative humidity at each time step could be obtained through linear interpolation. Finally, the aerosol optical depth (AOD) and radiation effect of each aerosol type (except for ammonium) can be calculated using BCC_RAD.

Data and Experiment Design
In this study, 1850 and 2010 (RCP4.5) were chosen to represent the status of the PI and PD, respectively. In November 2014, the U.S.-China Joint Presidential Statement on Climate Change highlighted the actions agreed by China and the United States to address climate change after 2020 and proposed to reduce greenhouse gas and other air pollutant emissions by 2030. Therefore, 2030 was chosen to represent the near-term future and 2050 to represent the middle-term future (MT) assessment years (Table 1).
Previous studies have found that oxynitride (NO x ) and sulfur dioxide (SO 2 ) are precursors of nitric acid (HNO 3 ) and sulfuric acid (H 2 SO 4 ), respectively. H 2 SO 4 competes with HNO 3 to react with ammonia (NH 3 ), which affects the formation of nitrate aerosols (Bauer et al., 2007;Bellouin et al., 2011;Bian et al., 2017;Wang et al., 2006). The principal gases that are related to the formation of nitrate aerosols in the atmospheric chemical module are NO x , NH 3 , and SO 2 ; therefore emission data for these gases from the RCP database Version 2.0 (Moss et al., 2010;van Vuuren et al., 2011) were used in the model. Figure 2 shows the time series of global and annual averaged emissions of NO x , SO 2 , and NH 3 . Both NO x and SO 2 emissions are declining since 2010 (except for NO x under RCP8.5); however, the emission of NH 3 shows an increasing trend until 2050. To investigate the variation in nitrate for the different emission scenarios, emission data from RCP4.5 for 2010 and RCP4.5, RCP6.0, and RCP8.5 for 2030 and 2050 were used. The DRF and ERFari values of nitrates were estimated as the differences between a specified year and the PI. Based on the emission data of the period, calculations were performed using the BCC_AGCM2.0_CUACE2.0. Sea surface temperature (SST) and sea ice cover (SI) were fixed as 21-yr climatology  from the Hadley Centre during the calculations (Hurrell et al., 2008), and the model integration time was 30 years (Pincus et al., 2016).
Observations of nitrate surface concentrations were provided by the Clean Air Status and Trends Network (CASTNET), European Monitoring and Evaluation Programme (EMEP), and China Meteorological Administration Atmosphere Watch Network (CAWNET) (Zhang, Wang, Niu, et al., 2012), which cover North America, Europe, and some parts of China, respectively. CASTNET, EMEP, and CAWNET sites used for validation provided weekly mean data for 2010, daily mean data for 2010, and monthly mean data from 2006 to 2007, respectively. There were 70 CASTNET, 28 EMEP, and 14 CAWNET sites used in this study. Observations of the total AOD at 0.55 μm were obtained from the MODIS Adaptive Processing System (MODAPS), and the monthly global data collected from the Terra platform (MOD08_M3, Ver.6.0) at 1°× 1°resolution from 2008 to 2012 were analyzed.
In this work, we used a scaling method to make the simulated surface concentration results of nitrate match the observational results well in China region. First, we estimated a ratio based on the simulated and the observational values, and then we obtained a scaling factor of 3 to enlarge the RCP source emission value in China region. To calculate DRF values, the "double radiation call" method was used, which calls the radiation scheme twice in the same radiation time step. The model meteorological field did not change when the radiation scheme was called twice. In the first call, all the aerosols were included when calculating the radiative forcing. In the second call, the calculated radiative forcing did not contain nitrates. During the same time step, aerosols affected only the radiative process without affecting other climate processes. The difference in radiative flux at the TOA during the two calls was the radiative forcing due to nitrates, and the difference in radiative forcing between two different time slices was the DRF during a specified period due to nitrate. The Radiative Forcing Model Intercomparison Project states that model ERF can be diagnosed by suppressing response, that is, specifying SST and sea ice concentrations   (Pincus et al., 2016). So, in this work, ERF values were calculated using the method proposed by Hansen et al. (2005); that is, by fixing SSTs and SI cover at climatological values while allowing all other parts of the system to respond until reaching a steady state in which the difference in net radiative flux at TOA was the ERF.

Surface Nitrate Concentrations
Figure 3 presents scatterplots that compare annual mean nitrate and sulfate surface concentration observations with simulations. They show that the simulated surface concentrations agreed reasonably well with the observed concentrations in North America, Europe, and China. The simulated values were within a factor of 2 of the observations at many of the sites. However, simulated values were lower than measured values at some sites where high concentrations were observed. This was likely caused by the interpolation method

10.1029/2019MS001622
Journal of Advances in Modeling Earth Systems used in this work and the accuracy of the observation data whose deviation was too large. The poor spatial results of observations in China and the limitations of observational conditions are also important reasons for the underestimates in China. Observation sites may be representative of regions no larger than a few kilometers (Hodnebrog et al., 2011;Stock et al., 2014;Wild & Prather, 2006). The resolution of the model is low; as Young et al. (2018) pointed out, spatial resolution can be a particular problem for comparison of global models with observations in urban regions, where strong precursor sources lead to large variations in

Journal of Advances in Modeling Earth Systems
nitrates on spatial scales far smaller than the model grid scale. When the simulated results are interpolated to specific sites, it may cause underestimate (Bian et al., 2017;Hauglustaine et al., 2014;Liu & Liao, 2017).
We further evaluated the performance of the model for the seasonal cycle of nitrate surface concentrations in North America, Europe, and China (see Figures 4, 5, and 6). The observations showed that the maximum nitrate surface concentrations occurred during the winter and the simulations of the model matched the seasonal evolution shown in the observations. Simulated values across North America and Europe were very

Journal of Advances in Modeling Earth Systems
China based on emission inventory described in Streets et al., 2003. This indicates that extensive efforts are still required to improve simulations of aerosol characteristics in China in future studies (Li et al., 2015). In addition, seasonal variation in emissions, the locations of the observation sites, and the meteorological field all impact the differences between simulated and observed values. Figure 7 shows the global distribution of the total AOD at 0.55 μm derived from simulations and MODIS observations and the difference due to nitrates. The BCC_AGCM_CUACE2.0 reproduced the pattern of AOD; the global mean value was closer to satellite observations when nitrate aerosols were considered and the simulations across East Asia were also improved ( Figure 7c). The high AOD in the Sahara region of Africa was mainly due to high dust emissions. Inaccurate emissions, model resolution, and particle size distribution of aerosols may have caused the lower AOD simulated in Northern India and Eastern China.

Total Aerosol Optical Depth
In addition, calculation of AOD in the model does not consider ammonium salts and secondary organic aerosols, which lead to lower AOD values. Some studies have also shown that many climate models tend to underestimate AOD in Eastern China Shindell et al., 2013). Figure 8a shows the global annual distributions of nitrate loading for the PD relative to the PI and the global annual loading of nitrate aerosol increased by 1.50 mg m −2 from the PI to the PD. The greatest increase in It should be noted that maximum values over East Asia and South Asia exceeded 11 mg m −2 , much higher than the global average. This finding is consistent with observations reported by Malm et al. (2004) and Putaud et al. (2004), which indicated that high nitrate aerosol loading occurred mainly in highly developed industrial areas. Strong loadings in Central Africa, Malaysia, and nearby areas were mainly due to agricultural ammonia and biomass burning (Hauglustaine et al., 2014). The same results were also found from the perspective of zonal distribution. Figure 8b shows that nitrates were mainly distributed around 10-60°N, where most of the global population is located. And according to the IPCC Fourth Assessment Report (Solomon, 2007), 87% of the anthropogenic nitrate aerosols were from the Northern Hemisphere. The vertical distribution of nitrate aerosols showed that they were mainly concentrated at altitudes <4 km and the concentration decreased rapidly with height, indicating that chemical formation near the emission source was the main source of nitrate aerosols, consistent with the findings of Zhang (1996) and Bellouin et al. (2011).

Changes in Nitrate Loading
With regard to the seasonal evolution, the loading of nitrate for the PD was 1.73 mg m −2 during December, January, and February (DJF), mainly distributed in Central Africa, Europe, East Asia, and South Asia, and decreased to 1.20 mg m −2 (~30% lower) during June, July, and August (JJA; see Figure 9), with a clear

10.1029/2019MS001622
Journal of Advances in Modeling Earth Systems reduction in East Asia. This is because nitrate loading is affected not only by precursor emissions but also by temperature. Changes in temperature have a significant impact on nitrate aerosol formation through gasparticle equilibria, which favors colder temperatures (Bellouin et al., 2011;Liao et al., 2006;Pye et al., 2009;Racherla & Adams, 2006). Most high nitrate emission areas are in the Northern Hemisphere, where DJF and JJA correspond to the winter and summer, respectively. Temperatures are high in summer, and nitrates are easily decomposed by heat, decreasing loading. In addition, nitrate loading is also dependent on the solar zenith angle (i.e., photochemical conversion rate), relative humidity (i.e., gas-liquid equilibrium), and wet deposition Wang et al., 2006). Excess precipitation in East Asia and

Journal of Advances in Modeling Earth Systems
Malaysia can make wet deposition of nitrate more effective in summer, which also significantly reduces nitrate loading.
Atmospheric loading of nitrate aerosols is predicted to decline in the future under the RCP4.5, RCP6.0, and RCP8.5 scenarios (Figures 10 and 11), consistent with the results of Hauglustaine et al. (2014)   emissions are predicted to remain high in the future. However, it can be seen from Figure 11 that the nitrate loading will increase in South Asia and decrease in East Asia and Europe under RCP4.5 and 8.5 in the future, indicating that nitrate may play a more important role in South Asia. Figure 12 shows the time evolution of nitrate loading in the three major industrialized regions of East Asia, Europe, and South Asia. The PD nitrate loading in EA (9.60 mg m −2 ) is much higher than those in EU (4.58 mg m −2 ) and SA (3.47 mg m −2 ), but it shows a rapid decline in the future until 2050 (except RCP6.0). Under the RCP4.5 scenario, nitrate loading in East Asia is predicted to decline significantly due to implementation of proposed mitigation measures, reducing to 2.81 mg m −2 by 2050, a decrease of 70.7% relative to the PD. This is consistent with the precursor emissions (NO x and NH 3 ) reduction ( Figure 2). For RCP8.5, nitrate loading over East Asia is predicted to be 5.96 mg m −2 (reduce by 37.9%) and 4.37 mg m −2 (reduce by 54.5%) in 2030 and 2050 relative to PD, respectively. It should be noted that the worst-case predicted scenario for East Asia in 2050 is RCP6.0 and the nitrate loading over East Asia is 5.06 mg m −2 , with a growth of~10% relative to 2030 values. Since the projected nitrate loading is much higher in East Asia than in Europe and South Asia under RCP6.0 and RCP8.5 in the future, East Asia is likely to become the most important global nitrate emission source; therefore, more research should be conducted into atmospheric nitrates within East Asia. However, it is worth noting that nitrate loading in South Asia showed an increasing trend from 2030 to 2050 under the RCP4.5 and 6.0 scenarios. The nitrate loading in South Asia is predicted to be 3.23 mg m −2 by 2050 under RCP4.5, exceeding the nitrate loading in East Asia (2.81 mg m −2 ) and Europe (2.69 mg m −2 ), making this region the largest contributor to nitrate emissions in this scenario, mainly due to increased NO x emissions from local industry and increased agricultural NH 3 emissions ( Figure 2) (Bian et al., 2017;Boucher et al., 2013;Clarke et al., 2007;Smith & Wigley, 2006;Wise et al., 2009). Nitrate loading in Europe is predicted to decline under three RCP scenarios, also consistent with Hauglustaine et al. (2014), indicating that future reductions in pollutant emissions play an important role in changes in nitrate loading. Figure 13 presents the global distribution of nitrate DRF values at TOA for the PD relative to the PI. The global mean annual all-sky DRF value of nitrates is −0.14 W m −2 , close to predictions reported in IPCC AR5 (−0.11 W m −2 ) (Boucher et al., 2013). Table 2 compares calculated nitrate all-sky DRF in this work with other studies. Hauglustaine et al. (2014) highlighted that accumulation mode particles largely dominate  nitrate forcing (~90%). The nitrate loading simulated in Li et al. (2015) was pretty low, and this may be the major reason for their low DRF prediction (−0.025 W m −2 ). In addition, this study did not consider the contribution of fine-mode (<0.05 μm) nitrate extinction, which can also cause the underestimate of nitrate DRF. The spatial distribution of nitrate DRF values correlates with the distribution of its loading; high values appear over East Asia, Europe, and South Asia (−0.87, −0.35, and −0.33 W m −2 , respectively). The global mean annual clear-sky DRF of nitrates is −0.32 W m −2 , and the values over East Asia, Europe, and South Asia are −2.09, −1.07, and −0.64 W m −2 , respectively. The all-sky DRF value is much weaker than the clear-sky DRF value, mainly due to the blocking effect of clouds (Takemura et al., 2002;, which can block shortwave radiation scattering via nitrate aerosols, reducing the outgoing shortwave radiation at the TOA remarkably. It should be noted that the all-sky DRF values of nitrates over Europe are projected to be −0.17 to −0.23 W m −2 by 2050 for the three RCP scenarios. Although the increase in nitrate loading is higher over Europe (2.43-3.31 mg m −2 ) than over South Asia (2.28-2.84 mg m −2 ) from the PI to 2050, the all-sky DRF values of nitrates are predicted to be higher over South Asia (−0.24 to −0.29 W m −2 ) than over Europe. This is mainly due to the high levels of cloud cover over Europe, which block a large amount of solar radiation scattering by nitrate aerosols, resulting in lower negative all-sky DRF values over Europe. The clear-sky DRF values of nitrates are much higher over Europe (−0.54 to −0.75 W m −2 ) than over South Asia (−0.47 to −0.56 W m −2 ), and their spatial distribution matches the distribution of nitrate loading. Also, relative humidity, nitrate particle size distribution, atmospheric vertical distribution, solar zenith angle, surface albedo, and other cloud microphysical factors (such as the effective radius of cloud droplets) can also affect nitrate DRF values (Wang et al., 2006).

Nitrate Direct Radiative Forcing
Without considering the influence of clouds, the global mean annual clear-sky DRF value of nitrates was found to be double the all-sky DRF value, which may reach −0.22 to −0.29 W m −2 by 2030 and −0.20 to −0.26 W m −2 by 2050 ( Figure 15). The highest clear-sky DRF values of nitrates over three industrialized regions occur in 2030 in the RCP8.5 scenario. The regional mean annual clear-sky DRF value of nitrates over East Asia is expected to be −1.36 W m −2 , with a maximum of −2.61 W m −2 , which is higher than over South Asia (−0.63 W m −2 ) and Europe (−0.95 W m −2 ). It should be noted that clear-sky DRF values for RCP6.0 are predicted to be higher over East Asia and South Asia in 2050 than in 2030 because of increased ammonia emissions due to agricultural fertilizers. In addition, a decrease in sulfur dioxide will result in a reduction in sulfate, which also contributes to additional formation of nitrate aerosols (Boucher et al., 2013).

Nitrate Effective Radiative Forcing
IPCC AR5 (Boucher et al., 2013) proposes the concept for ERF, which is the change in net TOA downward radiative flux after adjusting for atmospheric temperature, water vapor, and cloud, but with surface This study −0.14 1850-2010 Li et al., 2015−0.025 1850-2000Hauglustaine et al., 2014−0.056 1850-2000Boucher et al., 2013 to −0.03) 1750-2011 Myhre et al., 2013−0.12 to −0.02 1850-2000) Xu & Penner, 2012−0.12 1850Bellouin et al., 2011−0.12 1860-2000Bauer et al., 2007−0.06 1750-2000Liao & Seinfeld, 2005−0.10 1800-2000 temperature or a portion of surface conditions remaining unchanged. The ERF is evenly distributed with no obvious patterns (not closely correlated with the change in nitrate loading) and is positive in many places; thus, the discussion focuses on global mean annual values (Zhang et al., 2018). Although there are multiple methods to calculate ERF values, SSTs and SI cover are usually fixed at climatological values unless otherwise specified (Pincus et al., 2016). Land surface properties (i.e., temperature, snow and ice cover, and vegetation) can change in this method. Calculation of ERF values requires more complex methods than calculation of DRF values, but the inclusion of additional rapid adjustments makes ERF a better indicator of the eventual global mean temperature response, especially for aerosols (Boucher et al., 2013).

Conclusions
In this study, a next-generation atmospheric aerosol/chemistry-climate model, BCC_AGCM_CUACE2.0, was developed, and the optical properties (i.e., extinction cross section, single scattering albedo, and asymmetry factor) of nitrate aerosols were calculated. The variations in nitrate aerosol loading, as well as PD and future DRF and ERF (as we didn't consider the aerosol-cloud interactions, the ERF refers to ERFari in this work) values under three different emission scenarios (RCP4.5, 6.0, and 8.5), were simulated. The model reproduced the geographical distribution and seasonal changes in nitrate concentrations well. However, biases might be introduced by the relatively low model resolution and the inaccuracy of emission inventory in high-emission areas when comparing with the observations after interpolation at some sites.
The global mean annual loading of nitrates was predicted to increase by 1.50 mg m −2 from the PI to the PD, with the largest increase occurring over East Asia (9.44 mg m −2 ), followed by Europe (4.36 mg m −2 ) and South Asia (3.09 mg m −2 ). In terms of seasonal evolution, it was predicted that high nitrate loading would occur during winter (DJF) with low concentrations in summer (JJA). The nitrate loading of the atmosphere was predicted to increase by 1.03-1.37 mg m −2 and 0.96-1.24 mg m −2 from the PI to 2030 and 2050, respectively. Similarly, the area with the largest increase in nitrate loading was predicted to be East Asia, with values of 4.03-5.79 mg m −2 in 2030 and 2.65-4.90 mg m −2 in 2050.
The present global mean annual all-sky nitrate DRF (PD-PI) value at the TOA was estimated to be −0.14 W m −2 . Stronger nitrate DRF was predicted to occur over industrialized regions in the Northern Hemisphere, especially in East Asia (−0.87 W m −2 ). By 2030, the all-sky DRF value was projected to be about −0.10, −0.09, and −0.12 W m −2 for RCP4.5, 6.0, and 8.5, respectively. By 2050, the all-sky DRF value was predicted to be about −0.09, −0.09, and −0.11 W m −2 for RCP4.5, 6.0, and 8.5, respectively. The strongest nitrate DRF was still predicted to occur over East Asia, while DRF was predicted to be stronger over South Asia than over Europe due to reduced cloud cover. Despite the effects of cloud cover, surface albedo also affects nitrate DRF values, and low surface albedo would result in stronger negative nitrate DRF. Therefore, the nitrate DRF was predicted to be stronger over the sea than over land. In addition, RH and wet deposition play important roles in determining the geographical distribution of nitrate DRF.
The present global mean annual ERF for nitrates was predicted to be −0.