1 Introduction

As important air pollution components and short-lived radiative forcing agents, atmospheric aerosols have significant influences on air quality, regional climate changes and human heath (IPCC 2013). Their loadings have increased significantly since preindustrial times. Studies have indicated that aerosols could affect the radiative energy balance of earth-atmosphere systems because they could substantially absorb or scatter shortwave solar radiation and act as cloud condensation nuclei (Twomey 1974; Albrecht 1989; Forster et al. 2007), subsequently altering the thermal dynamics characteristics of the atmosphere and hydrologic cycle (e.g., Zhuang et al. 2018a; Chen et al. 2020). The East Asian region is always affected by active winter or summer monsoons. Both East Asian summer or winter monsoon (EASM or EAWM) anomalies might induce extreme weather and climate events (Ding and Chan 2005; Wang et al. 2008), thus threatening human life, property and production safeties in this region and causing much worse air qualities (Zhu et al. 2012; Mao et al. 2017). Additionally, with rapid increases in populations and economies, this region has also suffered from serious air pollution frequently due to very high levels of aerosols and their precursor emissions in recent years (Zhang et al. 2009; Li et al. 2017; Chen et al. 2019). High levels of aerosols associated with a variety of compositions in East Asia (Zhang et al. 2012) could induce climate changes in different ways in this region due to their direct, semidirect and indirect effects (Zhuang et al. 2013a, b; Zhou et al. 2014; Chen et al. 2020), which could further affect the development of East Asian monsoons (Wang et al. 2015; Zhuang et al. 2018a). All of the above results imply that anthropogenic aerosols have close interactions with monsoon climates in East Asia.

Many studies on aerosol direct radiative forcing (DRF), climate effects, and aerosols interacting with East Asian monsoons have been performed in the past two decades (Menon et al. 2002; Liao and Seinfeld 2005; Wang et al. 2010; Sun et al. 2012; IPCC 2013; Wang et al. 2015; Zhuang et al. 2018a, b). However, the fourth and fifth IPCC (Forster et al. 2007 and IPCC 2013) reports showed that the global mean simulated aerosol DRF ranges from − 0.63 to + 0.04 W/m2 and from − 0.6 to − 0.13 W/m2, respectively, suggesting that large uncertainties in the aerosol DRF still exist. The IPCC (2013) further indicated that the species breakdown of the aerosol DRF is less certain than the total forcing. The global mean black carbon (BC) aerosol DRF ranges from + 0.05 to + 0.8 W/m2. At the regional scale, the uncertainties in the aerosol DRF are larger than those at the global scale (Zhuang et al. 2018b). For example, Zhuang et al. (2013a) suggested that the aerosol DRF varied from − 1.11 to + 0.45 W/m2 only due to the uncertainty of the aerosol mixing states. The aerosol DRF uncertainties are possibly derived from aerosol emissions, mixtures, profiles and so on (Forster et al. 2007). The aerosol mixing state is one of the key factors affecting DRF assessment, especially when there is BC in the atmosphere. Jacobson (2001) emphasized that the absorption of aged or internally mixed BC could be twice as large as that of fresh BC, which could make BC contribute importantly to large positive radiative forcing (Peng et al. 2016). Therefore, it is necessary to consider the aerosol mixing states to better understand the aerosol DRF and climate effects. Nevertheless, there have been only a few studies (e.g., Zhang et al. 2011; Zhuang et al. 2013a; Curci et al. 2019) having addressed this issue. Due to direct or indirect aerosol perturbations of solar shortwave radiation, the global or regional climate will change. The direct effect of scattering aerosols (e.g., sulfate and nitrate) could induce a decrease in near-surface air temperature and total precipitation due to its negative radiative forcing at both the top of the atmosphere (TOA) and the surface (Kiehl and Briegleb 1993; Wang et al. 2010). Therefore, the warming effect of CO2 could be considerably offset in some regions. In contrast, absorbing aerosols (e.g., dust, BC) could induce a more unstable atmosphere, result in regional warming and flooding, and play an essential role in pollution development (Menon et al. 2002; Zhuang et al. 2013b, 2019; Peng et al. 2016; Ding et al. 2016). Moreover, BC absorption would intensify regional- to global-scale warming, and its warming effect is thought to be surpassed only by CO2 (Jacobson, 2002). Although BC loading only accounts for a small part of the total aerosols (Zhang et al. 2012), it could effectively alleviate the cooling effect of the scattering aerosols (Zhuang et al. 2013a). Unfortunately, how much cooling would be offset is still unclear because of the lack of observations about aerosol mixings.

Due to the high level of air pollutant emissions (Li et al. 2017), aerosol DRF in Asian regions is much stronger, and its climate effect might be more important than that in most other regions. Thus, considerable research on the Asian aerosol affecting (affected) or interacting with East Asian monsoons (EAMs) has also been carried out, instead of just focusing on its radiative forcing and climate effect (Huang et al. 2007; Wu et al. 2008; Meehl et al. 2008; Sun et al. 2012; Jiang et al. 2013; Wang et al. 2015; Zhuang et al. 2018a, b; etc.). Lau et al. (2006) indicated that decreases in rainfall in East Asia might be related to an increase in absorbing aerosols over south Asia. Liu et al. (2009) and Sun et al. (2012) showed that East Asian summer monsoon circulation could be weakened due to direct effects of sulfate and dust aerosols. In contrast, black carbon showed different influences. Owing to the absorption of BC, the EAM circulation was weakened during winter but enhanced during summer, and precipitation was likely increased in southern China during summer (Zhuang et al. 2018a). Therefore, the warming effect of BC might considerably offset the cooling effects of scattering aerosols or the indirect effects of aerosols (Wang et al. 2015) on the development of EAMs. Additionally, both Wang et al. (2015) and Zhuang et al. (2018a) also investigated and quantified the responses of the aerosols themselves to the EAM anomalies. They further presented the importance of monsoon climate changes in affecting air pollution, as mentioned previously or implied in Liu et al. (2011). Similarly, uncertainties of the interaction between aerosols and EAMs still exist (Wu et al. 2015), despite relevant studies being performed in recent years.

Previous studies have indicated that aerosols in East Asia are important in regional climate change and air pollution issues. It is worth noting that most of them mainly focused on the individual or integrated aerosol effects instead of addressing the uncertainties of these effects. As stated above, the aerosol mixing state is one of the key factors affecting aerosol radiative forcing estimation (Zhuang et al. 2013a; Curci et al. 2019) and its climate effect assessment. Chýlek et al. (1995) indicated that aerosol mixing was very important when the average ratio of BC to sulfate exceeded 5%. Therefore, it is necessary to take this factor into account in Asian regions because BC accounts for more than 10% of the total scattering aerosols in most regions (Zhang et al. 2012). To determine the influence of aerosol mixtures on aerosol climate effects, this study investigates the interactions between direct anthropogenic aerosol effects and the East Asian summer monsoon when aerosols are mixed in different ways. A well-developed online coupled regional climate and chemistry modeling system (RegCCMS, Wang et al. 2010, 2015) is used here based on its good performance in simulating aerosol loadings, optical properties, and regional climate variations in East Asia. More importantly, the regional model has a stronger ability to reproduce realistic small-scale features of climates than the global model (Denis et al. 2002).

The methods are described in next section. The simulated results and discussions are shown in Sect. 3. Finally, the conclusions of this study are provided in Sect. 4.

2 Methodologies

2.1 Description of the regional climate and chemistry modeling system RegCCMS

RegCCMS is a well-developed online coupled regional climate and chemistry modeling system. In RegCCMS, the meteorological changes induced by aerosols are allowed to immediately feedback to the variations in air pollutants (Wang et al. 2004). Both direct and indirect aerosol climate effects can be investigated by this modeling system (Wang et al. 2010). It is composed of a regional climate model version 3 (RegCM3) (Giorgi et al. 2002) and a tropospheric atmosphere chemistry model (TACM) (Wang et al. 2003, 2010, 2015; Li et al. 2009). RegCM3 is a mature regional climate model that is based on Version 3 of the Penn State/NCAR Mesoscale Model MM5 (Grell et al. 1993) and further developed by the Abdus Salam International Centre for Theoretical Physics (ICTP) in Italy. It has been widely used to study the regional climate changes in East Asia and other regions (Giorgi et al. 2002; Qian et al. 2003; Wu and Fu 2005; Huang et al. 2007; Li et al. 2009). To estimate the aerosol direct effect, a radiative transfer mode based on the parameterization of the National Center for Atmospheric Research (NCAR) Community Climate Model (CCM3) is included in RegCM3 (Kiehl et al. 1996). The solar component of the radiative transfer scheme follows the δ-Eddington approximation of Kiehl et al. (1996), which includes 19 spectral intervals from 0.2 to 5 μm. The Grell cumulus convective precipitation scheme (Grell 1993) and BATS land surface scheme (Dickinson et al. 1993) is used here. These two schemes are recommended to be used in combination (Ajay et al. 2019). However, this model could not address the chemical processes of scattering aerosols (e.g., nitrate). Thus, it is coupled online with TACM, in which a thermodynamic equilibrium model ISORPOPIA (Fountoukis and Nenes 2007) is included to simulate sulfate and nitrate aerosols. The mechanism of gas phase chemistry is derived from CBM4. Wet depositions of air pollutants are parameterized as a function of rainfall intensity. More details in the TACM description can be found in Wang et al. (2010).

The aerosol specific extinction coefficient, single scattering albedo and asymmetry factor are the key optical parameters in determining the aerosol direct radiative forcing. For externally mixed aerosols, the optical parameter of each type of aerosol is calculated individually. For example, the extinction coefficient (τ) of carbonaceous aerosols can be expressed according to Kasten (1969):

$$\tau_{i} \left( \lambda \right) = M_{i} \beta_{\lambda i} \left( {1 - RH} \right)^{{ - \kappa_{i} }}$$
(1)

where λ and RH are the wavelength and relative humidity, respectively. i represents four aerosol species, including hydrophilic BC and organic carbon (OC) and hydrophobic BC and OC. Mi is the mass of aerosol i. βλi is the wavelength-dependent specific extinction coefficient (m2/g) of aerosol i. κi = 0 for hydrophobic carbonaceous aerosols and equals 0.25 and 0.2 for hydrophilic BC and OC, respectively. βλi is a function of the aerosol refractive indices and physical properties (Qian et al. 2003). For sulfate and nitrate aerosols, their extinction coefficients can be expressed as follows based on Kiehl and Briegleb (1993) and Kiehl et al. (2000):

$$\tau_{i} \left( \lambda \right) = M_{i} \beta_{\lambda i,0} e^{{\left( {\beta_{\lambda i,1} + \frac{{\beta_{\lambda i,2} }}{{RH + \beta_{\lambda i,3} }} + \frac{{\beta_{\lambda i,4} }}{{RH + \beta_{\lambda i,5} }}} \right)}}$$
(2)

where βλi,0 is the wavelength-dependent specific extinction coefficient of aerosol i. βλi,1~5 are wavelength-dependent fitting coefficients of βλi,0 (Kiehl et al. 2000). The total extinction coefficient of aerosols is the sum of each aerosol.

For internally mixed aerosols, aerosols are mostly mixed in a core-shelled way: hydrophilic BC acts as a core surrounded by scattering aerosols (Jacobson 2000). The optical properties (e.g., specific extinction coefficient βλi,int) of the mixed particle are related to the volume fraction fvol of each aerosol i. Detailed introduction on the calculation of the internally mixed aerosol optical property can be referred to Yu et al. (2013). They pointed out that there are several steps to obtain the internal mixed aerosol optical properties. First, the volume fraction of the core and shell (Worringen et al. 2008) should be calculated based on fvol of each aerosol. Then the radius of the core (rc) and shell (rs) are obtained by solving Eq. (2) in Yu et al. (2013). Third, two parameters relative to rc, rc, and λ are input to the core–shell mode algorithm developed by Yang (2003) to calculate the optical properties of a multi-layered sphere. Finally, a database of optical properties (e.g.: βλi,int) of aerosols is created for internal mixing aerosols, which is relative to the volume fraction fvol of each aerosol. For example, the extinction coefficient of internally mixed aerosol (τi,int) can be expressed as:

$$\tau_{{i,{\text{int}}}} \left( \lambda \right) = \beta_{{\lambda i,{\text{int}}}} \left( {f_{vol,sulfate + nitrate} ,f_{vol,BC} ,f_{vol,OC} ,f_{vol,dust} } \right)\sum M_{i}$$
(3)

Three types of aerosol mixing states are considered in this study, including external (EM), internal (IM) and partial internal (PIM) mixing states. In addition, only anthropogenic aerosols are taken into account, meaning that the fvol of dust aerosols equals 0. Aerosols are all distinct from each other in EM, and τ is the sum of each aerosol, as mentioned above. All hydrophilic aerosols, which are distinct from hydrophobic carbonaceous aerosols, are internally mixed in IM. τ in IM is the sum of internally mixed hydrophilic aerosols and hydrophobic carbonaceous aerosols. Both EM and IM are thought to be extreme cases in the atmosphere, and aerosols are likely partially mixed (Zhuang et al. 2013a; Curci et al. 2019). In PIM, only a portion of hydrophilic aerosols are internally mixed, and the rest are externally mixed. As a result, τ in PIM is the sum of internally mixed hydrophilic aerosols’, externally mixed hydrophilic aerosols’ and hydrophobic carbonaceous aerosols’. It’s worth noting that the time or location dependent aerosol mixing ratios are still very difficult to be obtained currently due to lack of systemically observations as indicated by Zhou et al. (2018). Although some observational data or researches have showed that the internally mixed aerosol fractions are mostly varied from 10 to 50% (Adachi and Buseck 2008; Kim et al., 2008; Deboudt et al. 2010; Cappa et al. 2012; Zhuang et al. 2018b; Curci et al. 2019), there still have a gap to fully understand the aerosol mixing issues. As a result, the aerosol internally mixed ratios are always set as a fixed value for all aerosols within the above range when investigating the aerosol climate effects. For example, Zhang et al. (2011) and Zhou et al. (2018) assumed that the internally mixed aerosol ratio was 50% (maximum) and 30% (middle), respectively. And they did not consider the difference among aerosol species. In contrast, Kim et al. (2008) applied a model predicted mixing ratio of each aerosol specie, which had been validated by some observations and was different from each other. Therefore, the internally mixed ratios of the aerosols proposed by Kim et al. (2008) are considered to be used in this study. There have approximately 32.2, 32.2, 35.5 and 48.5% of sulfate, nitrate, BC and OC, respectively, being assumed to be internally mixed.

2.2 Experiments

The model domain covers most regions of East Asia with a horizontal resolution of 50 km. Its center is set at (34.5° N, 106.8° E). In the vertical direction, there were 18 layers from the surface to 50 hPa. The National Centers for Environmental Prediction (NCEP) reanalysis data (NNRP2) with a 2.5° × 2.5° spatial resolution are employed as the initial and lateral boundary conditions during simulations. The weekly mean sea surface temperature (SST) is used by RegCCMS and kept as a fixed value when investigating aerosol climate effects. SST data are from the National Oceanic and Atmospheric Administration (NOAA) Optimum Interpolation (OI) SST V2 (Reynolds et al. 2002). Notably, the chemistry mode of RegCCMS would incur enormous computational costs. Due to the limitations of the calculations and storage. The coupling model was integrated for only four months from May to August every year from 1988 to 2011. The first integrated month (May) of each year was a spin-up time, which was long enough for the simulations according to Giorgi and Bi (2000). Only the simulated results within the region (100–145° E, 15–55° N) in summer (JJA) of these 21 years are analyzed here. To exclude the influence of the interannual variation in emissions, only the emission inventory of aerosols and trace gases in 2006, which is from the Intercontinental Chemical Transport Experiment-Phase B (Zhang et al. 2009), is applied. The major anthropogenic emissions of air pollutants have been estimated in Asian regions based on a series of improved methods (Zhang et al. 2009).

To understand the interactions between the direct aerosol effect and East Asian summer monsoon, as well as the influence of aerosol mixing states on the interactions, four numerical experiments were carried out, as shown in Table 1. Only sulfate, nitrate, black carbon and primary organic carbon (POC) aerosols from anthropogenic sources are taken into account in this study. Exp. C is the control run, which did not consider any effects of the aerosols. Experiments from Exp. S1 to S3 are the sensitive runs, which include the direct aerosol effect associated with external, internal and partial internal mixtures, respectively. The interactions between the different mixing aerosols and EASM are derived from Exp. S1–S3 minus Exp. C. To exclude the influence of the aerosol indirect effects, the aerosol affected cloud condensation nuclei (CCN) in the model is substituted by a background (prescribed) CCN value (Kristjánsson 2002) in both the control and sensitive experiments when addressing the cloud microphysics processes. Student’s t test method is applied based on seasonal mean simulations to test the statistical significance of the differences between sensitive experiments and the control experiment.

Table 1 Numerical experiment setup
Full size table

3 Results and discussion

3.1 Aerosol loadings, optical properties and radiative forcing

Both the aerosol and East Asian climate characteristics, as well as their variations, could well be simulated by RegCCMS, as indicated by the authors’ previous studies (Li et al. 2009; Wang et al. 2010, 2015; Zhuang et al. 2010, 2013a, b). Therefore, only brief comparisons between the simulated and observed aerosols and climates would be carried out here as shown in Fig. 1. The results to be analyzed in this section focus on the region within East Asia from 100° E to 145° E in the latitudinal direction and from 15° N to 55° N in the meridional direction. The simulated thermal-dynamic fields are compared against the NCEP reanalysis data. The simulated total precipitations are compared against the Tropical Rainfall Measuring Mission (TRMM) products. The simulated aerosol surface concentrations are compared against ground based observation (Zhang et al. 2012). In Zhang et al. (2012), there have 14 observation sites, including seven urban stations and seven regional stations, to collect aerosol samples and analysis their concentrations. These stations cover most parts of China from southern to northern, as well from western to eastern regions. Detailed introduction on the stations can be referred to Zhang et al. (2012). The simulated aerosol optical depths are compared with satellite-based Moderate Resolution Imaging Spectroradiometer (MODIS) data. Comparisons suggest that the coupled model can capture the characteristics of East Asian summer climates and aerosols. Overall, the spatial distributions of the air temperature, wind fields, as well as the aerosol concentrations and optical depths from simulations are likely consistent with those from Reanalysis products or observations. The simulated air temperature in continent regions in East Asia is lower than the Reanalysis data, thus the trough at 520 hPa is relatively stronger. In contrast, the simulated precipitations are less consistent with observations comparing with other simulated valuables. The total precipitation from simulation is smaller than that from TRMM product along the Yangtze River regions. Underestimations of the precipitations in Center to South China in summer also occurred in other simulations (Zhou et al. 2014; Yin et al. 2015; Sun and Liu 2016). Both the black carbon and total scattering aerosols (a sum of sulfate, nitrate and primary organic carbon) are well correlated to the ground observation. Linear correlation coefficients between the simulated and observed aerosol concentrations are 0.80 for BC and 0.79 for total scattering aerosols, which are all statistically significant at the 99% confidence level. The simulated aerosol optical depths shown in the figure are higher in southwest China than the MODIS AODs at 550 nm and they are more close to each other in the rest areas. It’s worth noting that the output of simulated AODs are at 495 nm, which are larger than those at 550 nm (Zhuang et al. 2014). Therefore, the simulated AODs at 550 nm are slightly smaller than MODIS in most region of East Asia except that in Southwest China.

Fig. 1
figure 1

Comparisons between the simulations and reanalysis or observed products, a Air temperature (K) and wind (m/s) at 520 hPa, the left one for RegCCMS and right one for NCEP, b Precipitation (mm/day), the left one for RegCCMS and right one for TRMM, c The aerosol surface concentration (μg/m3), the left one for RegCCMS and right one for comparisions, d The aerosol optical depth, the left one for RegCCMS and right one for MODIS

Full size image

Figure 1 further presents that all aerosols have high loading levels in Southwest and North China. They are more northerly in distribution in summer than in winter (not shown here) due to the influence of summer monsoon transportation (Liu et al. 2011). Sulfate aerosol, which accounts for more than half of the total aerosols considered in this study, is still the major component of anthropogenic aerosols in East Asian regions, followed by nitrate, POC and BC. The seasonal mean surface concentration of the total aerosols exceed 90 μg/m3 in Southwest China. In the area of 100–130° E, 25–45° N, BC accounts for approximately 8% of the total aerosols for the total column burden and 13% for the surface concentration, suggesting that the aerosol mixing states should be taken into account when assessing its direct radiative forcing and climate effect. This ratio is consistent with the observations in Zhang et al. (2012).

The aerosol optical and direct radiative forcings have considerable differences among the different aerosol mixing states. Taking the aerosol single scattering albedo (SSA) as an example, it will get smaller with increasing the proportion of internal mixing aerosols. Figure 2 shows the seasonal mean SSAs of the EM and PIM aerosols, as well as that from satellite-based Ozone Monitoring Instrument (OMI) product in summer. Similar to that in Zhuang et al. (2013a), the IM SSA, which is not shown here, is the smallest among these three mixing assumption. Due to the high level of BC in continental regions, the aerosol SSA is likely smaller in these areas than in other regions. The smallest SSA appears in central to North China, with values of approximately 0.92, 0.60 and 0.90 for EM, IM and PIM aerosols, respectively. Their regional means are 0.97, 0.70, and 0.95, respectively, showing a variation of 0.27 in summer due to the aerosol mixtures. This variation might further result in substantially different aerosol DRFs among different aerosol mixing states. Comparison between the simulated and OMI SSAs in the figure suggests that the PIM aerosol SSA is much closer to the observation, which is consistent with that in Zhuang et al. (2013a). The regional mean SSA from OMI is about 0.948, 0.006 smaller than that of PIM aerosols. Zhou et al. (2018) points out that the aerosol mixing is a complex issue and it’s hard to fully explore the impacts of internal mixing aerosol in current stage. Investigation here reflects to their statement to degrees. Therefore, comparison of SSAs might be not as good as that of the aerosol concentrations or optical depths.

Fig. 2
figure 2

The season mean aerosol single scattering albedo (SSA) of external (left, EM) and partial internal (middle, PIM) mixed aerosols, as well as that from OMI products (right) in East Asia in summer

Full size image

Meanwhile, both instantaneous and effective aerosol DRFs at the top of the atmosphere (TOA) are also investigated here. Differently from the instantaneous DRF, the latter takes regional climate feedback into account when assessing its value. They are derived from the differences between the sensitivity experiments and the control experiment. The results are plotted in Fig. 3. The IM aerosols can exert a very positive instantaneous DRF at the TOA in summer, with a maximum exceeding 5 W/m2. In contrast, the EM aerosols mainly exert a negative instantaneous DRF. However, the strongest instantaneous DRF of EM aerosols is not always found in areas with high aerosol concentrations because the negative forcing of scattering aerosols is substantially offset by BC. The negative instantaneous DRF becomes increasingly weaker with an increase in the proportion of IM aerosols. Thus, the instantaneous DRF of PIM aerosols is not obvious in continental regions compared with those of the other two types of mixed aerosols. The regional mean instantaneous DRFs of the EM, IM and PIM aerosols are − 0.95, 1.50 and − 0.12 W/m2, respectively, showing in Table 2. The aerosol mixtures mainly cause great differences in the TOA instantaneous DRFs of the aerosols. The surface instantaneous DRFs are all negative, and their spatial distributions (not shown here) are consistent with the aerosol concentrations. Moreover, their differences among EM, IM and PIM are much smaller than the instantaneous DRFs (Table 2). Different from the instantaneous DRFs, the EM, IM and PIM aerosols all exert negative effective DRFs at both the TOA and surface, implying that the influence of the mixing states on the aerosol TOA radiative forcing might be weaker than on instantaneous forcing due to the adjustment of the regional climate feedback. Their spatial distributions are more consistent with the aerosol loadings. Nevertheless, internally mixed aerosols can still significantly offset the negative effective DRF of aerosols (Fig. 3). Due to the absorption of BC or internally mixed aerosols, EM, IM and PIM all exert positive effective DRFs in the atmosphere (Table 2) in most areas of the region, especially in Southwest China. The more aerosols are internally mixed, the stronger the positive forcing in the atmosphere with larger coverage.

Fig. 3
figure 3

The instantaneous direct radiative forcing of aerosols under cloudy sky conditions (top panel, W/m2) and effective direct radiative forcing (bottom panel, W/m2) at the top of the atmosphere for external (left, EM), internal (middle, IM) and partial internal (right, PIM) mixed aerosols in East Asia in summer. The black dots in the bottom panel indicate the 90% confidence levels from Student’s t test

Full size image
Table 2 The regional mean aerosol direct radiative forcing (W/m2) over the region (100–130° E, 25–45° N) in summer
Full size table

Table 2 summarizes the regional mean aerosol DRFs in different mixing states at both the TOA and surface, as well as in the atmosphere in summer. This shows that the TOA forcing of instantaneous forcing is weaker than that of effective forcing, which is opposite to that at the surface. The strong negative instantaneous DRF is significantly contributed by the absorbing aerosols. The adjustment of the climate feedback seems to favor reducing the gap between the TOA and surface forcing, which was also found in Zhuang et al. (2018a). Due to the absorption of BC or internally mixed aerosols, the mean effective DRFs in the atmosphere are all positive despite the aerosol mixing states. Surface effective DRFs are more sensitive to aerosol mixings than instantaneous DRFs. The regional mean (100–130° E, 25–45° N) effective DRFs of the aerosols range from − 5.54 to − 3.60 W/m2 at the TOA and form + 0.42 W/m2 to + 2.76 W/m2 in the atmosphere. Overall, the aerosol mixtures will lead to a considerable variation in the TOA forcing, as 2.45 W/m2 for instantaneous DRF and 1.94 W/m2 for effective DRFs in the region in summer. The aerosol DRF in the East Asian region is most likely close to − 0.12 W/m2 for instantaneous DRF and − 5.16 W/m2 for effective DRF because the PIM aerosols are closer to the real conditions, as mentioned above or suggested by Zhang et al. (2011), Zhuang et al. (2013a), and Curci et al. (2019). Based on observation, Li et al. (2010) presented that cloud-free aerosol instantaneous radiative forcing across China was − 15.7, + 16, and + 0.3 W/m2 at the surface, within the atmosphere, and at the TOA. In this study, corresponding values of PIM aerosol in East Asian region are − 20.95, + 11.82, and − 9.13 W/m2, respectively. They (Li et al. 2016a, b) also indicated that the aerosol may cool the surface and warm the atmosphere considerably, which is also found here. The results in Table 2 further indicate that the regional aerosol DRFs are much stronger than the global DRFs, as are their uncertainties (Forster et al. 2007; IPCC 2013).

3.2 Effects of different mixed aerosols on the East Asian summer monsoon

Similar to the aerosol effective DRFs, all the results to be analyzed below are from the differences between the sensitive and control experiments. Once solar radiation is absorbed to different degrees, the shortwave heating rate (SHR) of the atmosphere is increased, especially in regions or layers with high BC or internally mixed aerosols. Figure 4 illustrates the changes in seasonal mean SHR in the vertical direction in summer. The SHR anomalies can be as large as 1.8 K/day in north China. Its positive anomalies may extend near 850 hPa. Similarly, aerosol mixings also play an important role in SHR anomalies. The more aerosols are mixed internally, the more solar radiation will be absorbed (Table 2); subsequently, more positive SHR anomalies will be found. Due to the effect of the aerosol mixing states, the SHR varies from 0.26 to 0.34 K/day in the region (100–130° E, 20–45° N) in summer. The regional change in SHR due to the effect of PIM aerosols is approximately 0.28 K/day. Although East Asian regions are dominated by scattering aerosols, the direct effect of anthropogenic aerosols on regional climate changes in this region might be different from that of scattering aerosols because of the existence of BC. Relevant results have been also found in Lu et al. (2020).

Fig. 4
figure 4

Changes in the vertical shortwave heating rates (K/day) in the altitude-latitude section averaged from 105° to 125° E due to the external (left, EM), internal (middle, IM) and partial internal (right, PIM) mixed aerosol direct effects over East Asia during summer. The black dots in the figure indicate the 90% confidence levels from Student’s t test

Full size image

Increases in SHR in the lower atmosphere further result in anomalies of the atmospheric thermal fields, as shown in Fig. 5, which presents the seasonal mean changes in low layer (or near surface) air temperatures from the surface to 850 hPa in the East Asian region in summer under conditions with different aerosol mixtures. Low layer air temperatures have substantial increases due to the absorption of aerosols. Most increases in air temperature below 850 hPa appear in southwest and north China, with a maximum value exceeding 0.4, 0.7, and 0.5 K, respectively, due to the effects of EM, IM and PIM aerosols in summer. As indicated in Sect. 3.1, aerosols are conductive to cooling surface due to its negative radiative forcing. Surface temperature is decreased almost in whole continent region (not shown here) due to the aerosol direct effect. However, this cooling effect could be partially offset by the reduction of low layer cloud amount as shown in Fig. 6. Aerosol warming in boundary layer is also affected by the surface cooling especially for the atmosphere closer to the surface, leading to the most warming layer appearing near 900 hPa as shown in Fig. 7. The regional mean (100–130° E, 20–45° N) near-surface air temperature changes range from 0.09 to 0.17 K due to the influence of aerosol mixing states in summer. PIM aerosols cause an increase of approximately 0.11 K. The regional mean surface air temperature changes vary only from 0.02 to 0.06 K.

Fig. 5
figure 5

Changes in the air temperatures (K) below 850 hPa due to the external (left, EM), internal (middle, IM) and partial internal (right, PIM) mixed aerosol direct effects over East Asia during summer. The black dots in the figure indicate the 90% confidence levels from Student’s t test

Full size image
Fig. 6
figure 6

Changes in the cloud cover (%) below 850 hPa due to the external (left, EM), internal (middle, IM) and partial internal (right, PIM) mixed aerosol direct effects over East Asia during summer. The black dots in the figure indicate the 90% confidence levels from Student’s t test

Full size image
Fig. 7
figure 7

Changes in the vertical air temperatures (shaded, K) and circulations (vectors) in the altitude-latitude section (top panel) averaged from 105° to 125° E, as well as the changes in 850 hPa wind fields (bottom panel: m/s) and specific humidity (bottom panel, shaded: g/kg) due to the external (left, EM), internal (middle, IM) and partial internal (right, PIM) mixed aerosol direct effects over East Asia during summer. The black and blue dots in the figure indicate the 90% confidence levels from Student’s t test

Full size image

An increase in air temperature will affect cloud formation on the one hand and induce dynamic anomalies of the atmosphere on the other hand. Figure 6 illustrates the changes in cloud amount (CA) in the lower atmosphere below 850 hPa in summer. This result indicates that the EM, IM and PIM aerosols all lead to reductions in CA in most regions of East Asia, with the largest change exceeding − 2.5%. The distributions of the CA anomalies are highly consistent with those of air temperature in summer, as shown by the comparison between Figures 5 and 6, which further implies that the CA changes here are dominated by the aerosol semidirect effect. This effect is also found in Zhuang et al. (2013b). Thus, the lower atmospheric clouds will likely decrease more significantly when aerosols are mixed internally. The reductions in CA favor surface warming, which might offset part of the cooling effects due to the negative aerosol DRFs at the surface. The regional mean CA changes induced by EM, IM and PIM aerosols in East Asia in summer range from 0.46 to 0.53%. It is worth noting that the CA changes are not only affected by the air temperature anomalies but also derived from the dynamic adjustment of the atmosphere.

Figure 7 shows the anomalies of atmospheric meridional circulations and wind fields at 850 hPa due to the influences of EM, IM and PIM aerosol direct effects in summer. Due to the direct effect of anthropogenic aerosols, there is a warming zone in the lower atmosphere in the middle latitudes, which induces an increase in southerlies near the surface and an enhancement in ascending motion. Additionally, cyclonic circulation anomalies appear in southwest and northeast China. Therefore, the direct effect of anthropogenic aerosols in East Asia enhances the East Asian summer monsoon circulation, which is similar to the BC warming effect on the EASM. The results here may further imply that the BC warming effect in East Asian regions is much stronger than that of the remaining anthropogenic aerosols, especially when BC is internally mixed with scattering aerosols. The dynamic anomalies in southwest to south China and in northeast China favor cloud formation, which offsets part of the CA reductions due to the aerosol warming effect in these regions (Fig. 6). This offsetting effect is more obvious if more aerosols are internally mixed with each other.

Total precipitation is inhibited in most areas (including Center, North and East China) of East Asia, possibly because of the reductions in CA due to the aerosol semidirect effect, as shown in Fig. 6. The warming effect of mixed aerosol (Fig. 5) is in favor of reducing cloud amount (Fig. 6) and the atmospheric circulation anomaly has a relative weaker influence on the atmospheric moisture transportations (Fig. 7) in these regions. Similar to the CA variations, the figure also indicates that the decreases in rainfall are slightly more significant in the Sichuan Basin region (30° N, 105° E) and North China than in other regions in East Asia in summer, with a change of approximately 0.3 mm/day. The surface will become dryer if the IM aerosol effect is taken into account. Although IM aerosols can cause a greater reduction in precipitation in some regions, as mentioned above, they can also induce a more serious (or less reduction) increase in rainfall in other regions. For example, the atmospheric circulation anomalies induced by the aerosol direct effect mainly appear in South China and northeast Asia, where the atmospheric moisture contents are increased to degrees (Fig. 7), especially when the aerosols are more mixed internally. And the precipitations in these regions have a positive anomaly. The increases in rainfall reach the most due to the influence of IM aerosols. PIM aerosols lead to changes in total precipitation in summer of − 0.04 mm/day on average in the region (100°–130° E, 20°–45° N). Overall, a more significant aerosol warming effect would bring a more enhancement in EASM circulation and a more obvious aerosol semidirect effect. And the precipitation changes in East Asia in summer are compositely affected by these factors.

Table 3 further summarizes some regional mean climate factor anomalies in the East Asian region in summer due to direct effects of anthropogenic aerosols with different mixing states. This shows that the direct effect of anthropogenic aerosols may result in regional warming and drying in most regions of East Asia in summer. Additionally, the EASM circulation is enhanced. All the sensitivity experiments obtain similar results. Thus, the coeffects of scattering and absorbing aerosols on the East Asian summer monsoon are more clear based on the results here. The direct effect of anthropogenic aerosols on the EASM is similar to that of BC (Zhuang et al. 2013b; 2018a). However, differences exist among different experiments, indicating that the more aerosols are mixed internally, the more significant responses of the East Asian climate will be found. Therefore, the responses of air temperature and wind speed become increasingly significant with an increasing internally mixed ratio of aerosols. In contrast, regional mean changes in precipitation and CA do not vary monotonically because their positive responses in some regions may substantially offset their negative responses, as analyzed above, and vice versa. The PIM aerosol direct effect results in an increase of 0.28 K for lower atmosphere temperature and by 0.12 m/s for wind speed at 850 hPa, as well as a decrease by 0.04 mm/day for total precipitation in the region (100°–130° E, 20°–45° N) in summer. Previous studies on aerosols affecting East Asian climates suggested that the BC warming effect could lead to variations in summer precipitation, with increasing rainfall in southern China and droughts in northern China over the last five decades (Menon et al. 2002). The sulfate direct effect and total aerosol combined effects, including the aerosol direct and indirect effects, might weaken the EASM, as suggested by Liu et al. (2009) and Wang et al. (2015), respectively. Li et al. (2016a, b) reported that the absorbing aerosol-radiation interaction could alter the land–ocean thermal contrast, the thermodynamic stability and convective potential of the low atmosphere, leading to a strengthening of the early monsoon and affecting the subsequent evolution of the monsoon. The results here are consistent with Menon et al. (2002), Li et al. (2016a, b) and Zhuang et al. (2018a). They further emphasize the importance of BC and its mixing states in affecting the East Asian climate.

Table 3 Regional (100°–130° E, 20°–45° N) mean variations of near surface SHR, air temperature, 850 hPa wind speed and precipitation due to the effects of mixed aerosols in summer
Full size table

3.3 Aerosol feedbacks to the responses of the East Asian summer monsoon

Aerosol loadings are highly determined by meteorological factors. Recent studies have also pointed out that the variations or anomalies of East Asian climates could lead to considerable changes in aerosol distributions and concentrations (Liu et al. 2011; Zhu et al. 2012; Mao et al. 2017; Zhuang et al. 2018a, etc.). Zhuang et al. (2018a) suggested that BC changes from its warming effect were as important as those resulting from EASM interannual variations. Analysis in the previous sections shows that the EASM has been affected by mixed aerosols. Thus, aerosol loadings have also been changed in sensitive experiments compared to those in the control run. Figure 9 presents the changes in the anthropogenic aerosol concentrations at the surface in summer due to the direct aerosol effect. All aerosols increased to a certain degree in most regions of East Asia due to their direct effect, despite the mixing states. The greatest increases are likely found in southwest, north and parts of northeast China, with variations ranging from 5 to 25 μg/m3. The aerosol responses are very sensitive to the precipitation and atmospheric circulation anomalies, as shown in Figs. 7, 8 and 9. The cyclonic circulation anomalies in southeastern and northeastern China lead to the accumulation of aerosols. Thus, the aerosol loading anomaly is also significant in Northeast China. Additionally, the wet scavenging of air pollutants is inhibited from Southwest China to North China, which results in more aerosols being concentrated in the atmosphere. In contrast, the increase in precipitation in South and Northeast China offset parts of the increases in aerosols, especially in South China. Changes in aerosol loadings are also sensitive to aerosol mixing states. Overall, the more aerosols are mixed internally, the greater their loadings will be increased, implying that aerosol mixtures might be important in affecting air pollution.

Fig. 8
figure 8

Changes in the total precipitation (mm/day) due to the external (left, EM), internal (middle, IM) and partial internal (right, PIM) mixed aerosol direct effects over East Asia during summer. The black dots in the figure indicate the 90% confidence levels from Student’s t test

Full size image
Fig. 9
figure 9

Changes in the total aerosol concentrations (μg/m3) at the surface due to the external (left, EM), internal (middle, IM) and partial internal (right, PIM) mixed aerosol direct effects over East Asia during summer. The black dots in the figure indicate the 90% confidence levels from Student’s t test

Full size image

For each aerosol component, its response to the EASM anomaly is similar to that of the total aerosol loadings. All of them are increased to different degrees due to the direct aerosol effect on the EASM, as shown in Fig. 10. Sulfate aerosol varied the most, followed by nitrate, POC and BC. Due to the effect of aerosol mixtures, the regional mean (100°–140° E, 20°–50° N) increase of total aerosol concentrations can be changed by 0.14 μg/m3, which is about 8.56% to the averaged increase of EM, IM and PIM considerations’. PIM aerosols can induce increases of 1.07, 0.20, 0.16 and 0.13 μg/m3 for sulfate, nitrate, POC and BC, respectively, in the region in summer. The total aerosol concentrations in this region in the PIM experiment is more than 11% higher than that in the control run. The responses of the aerosol loadings here are similar to those in Wang et al. (2015), in which aerosols also accumulated due to the aerosol combined effects. However, the incentives leading to the aerosol increase in Wang et al. (2015) are different from here.

Fig. 10
figure 10

Regional mean responses of the nitrate, sulfate, black carbon and primary organic carbon aerosol concentrations (μg/m3) at the surface over the region (100°–140° E, 20°–50° N) due to external (EM), internal (IM) and partial internal (PIM) mixed aerosol direct effects during summer

Full size image

3.4 Differences in the climate responses in strong and weak monsoon index years

The East Asia summer monsoon has significant interannual variations, which would result in different air pollutant distributions (Liu et al. 2011; Zhu et al. 2012; Mao et al. 2017; Li et al. 2018). Thus, the aerosol climate and environmental effects may vary in different years (Li et al. 2016a, b; Zhuang et al. 2018a). To propose a common monsoon index suitable for all known monsoon regions, Li and Zeng (2002) defined a dynamical normalized seasonality (DNS) index which was a function of climatological and monthly mean wind vectors at 850 hPa. And they have proved that DNS index can successfully describe the seasonal and interannual variability of the monsoons. Following the definition of the monsoon index in Li and Zeng (2002), Mao et al. (2017) indicated that there were five of the strongest and five of the weakest years for the EASM index during the simulated period. The five strongest EASM index (much more positive DNS anomaly) years are 1990, 1994, 1997, 2004, and 2006. The five weakest EASM index (much more negative DNS anomaly) years are 1988, 1993, 1995, 1996, and 1998. According to their classification criteria, the difference in the seasonal mean aerosol loadings and 850 hPa wind field in summer between the strongest and weakest monsoon index years are plotted in Fig. 11. There are substantial differences in the wind field at 850 hPa in East Asia, including in the northeast and southeast regions between the strongest and weakest ESAM index years, showing a more anomalous northeast wind in South to Southeast China and a northwest to west wind in Northeast China. These differences are highly consistent with the results in Mao et al. (2017) and Zhuang et al. (2018a), further implying the good ability of RegCCMS to simulate East Asian climates. In most regions of China, especially in southern China, all aerosol loadings are relatively higher in the strongest EASM index years than in the weakest years. The differences in sulfate, nitrate, POC and BC can reach 3.98, 0.77, 0.34, and 0.27 μg/m3 in the East Asia region between the strongest and weakest EASM index years. The aerosol changes from EASM variations are as important but are larger than those resulting from its climate effect, which is consistent with Zhuang et al. (2018a).

Fig. 11
figure 11

Differences in the surface aerosol concentrations (shaded, a and b for absolute changes in μg/m3 and c and d for relative changes in %) and in the wind fields at 850 hPa (d, arrow; m/s) between the strongest and weakest monsoon index years during summer. S, N, BC and POC in the figure represent nitrate, sulfate, black carbon and primary organic carbon aerosols, respectively. The purple and black dots indicate the 90% confidence levels from Student’s t test for wind fields and aerosol loadings, respectively

Full size image

Higher aerosol loadings in the strongest EASM index years over East Asia are mostly result from the differences of the meteorological factors between the strongest and the weakest monsoon index years. For example, higher aerosol concentrations in the strongest EASM index years in northeast and southeast areas of the region are likely caused by the influences of transportations. In addition to wind field, both the precipitation and air temperature also have substantial differences between the strongest and the weakest monsoon index years as shown in Fig. 12. In the strongest EASM index years, there have less precipitations in most areas of continent region while have heavier rainfalls in most parts of ocean than those in the weakest EASM index years. These differences between the strongest and weakest monsoon index years are consistent with the results in Li and Zeng (2002), in which the precipitations in summer had a negative correlation with EASM index in Center to South China while had a positive correlation on the ocean around South China Sea. Consequently, the wet scavenging of air pollutants is smaller in the region with decreasing precipitation, and vice versa. The pattern of the increases in aerosol loadings in continent regions are consistent with that of the precipitation anomalies to degrees, especially in central to east China where the wind differences are small, as shown in Figs. 11 and 12. The atmosphere at middle to low latitudes are much cooler in the strongest EASM index years than the weakest years, which are more conducive to the accumulation of air pollutants and the formation of secondary inorganic aerosols as indicated in Wang et al. (2015).

Fig. 12
figure 12

Differences in the precipitation (a, mm/day) and air temperature (b, K) below 850 hPa between the strongest and weakest monsoon index years during summer. The black dots indicate the 90% confidence levels from Student’s t test

Full size image

Under normal circumstances, sulfate and nitrate aerosols have a higher scavenging efficiency than primary carbonaceous aerosols. Furthermore, their formations are also affected by the air temperatures. Therefore, the relative increases of sulfate and nitrate in the strongest EASM index years are much larger than those of the primary carbonaceous aerosols in most areas of China (Fig. 11c, d). The surface concentrations can be 25–32% higher for sulfate and nitrate aerosols while only about 8.7% higher for the carbonaceous aerosols in the strongest EASM index years than in the weakest years in East Asia region [25–45° N, 100–130° E]. Subsequently, the climate responses to the mixed-dependent anthropogenic aerosol direct effect may be different between the strongest and weakest EASM index years. Taking the air temperature changes as an example (in Fig. 13), the warming effect of EM aerosols is substantially weaker in central to part of northern China in the strongest EASM index years than in the weakest years, with a difference exceeding − 0.5 K. This weakness of the warming effect is mostly derived from the changes in the aerosol loadings. As a result, the precipitation responses in the strongest EASM index years would be less positive or be more negative in most regions of South and Northeast China than in the weakest years. Changes in precipitation responses are mostly derived from the differences of the thermodynamic field adjustments between the strongest and weakest monsoon index years, which has been stated in Li et al. (2016a, b). In contrast, the weaknesses mentioned above are no more obvious, and considerable enhancements in the warming effect are found in north to the east China, as well as in parts of south China due to the IM aerosols in the strongest EASM index years. It is obvious that the internal aerosol mixings further strengthen the aerosol absorption and substantially offset parts of the cooling effect by the increased scattering aerosols. Subsequently, the precipitation responses in the strongest EASM index years would be more positive or be less negative in most regions of Center to South China, as well as in Northeast China. It worth noting that a significant increase in aerosol loading between the strongest and weakest monsoon index years are likely expected to induce a more increase in air temperature in the strongest monsoon index years as implied in Fig. 5. However, the relative increases of sulfate and nitrate aerosols in the strongest monsoon index years are at least 2–3 times larger than those of the absorbing carbonaceous aerosols as mentioned above. These additional increases in scattering aerosols then might offset the expected warming effect of the increased mixed aerosols to degrees. Overall, the warming effects due to the EM, IM and PIM aerosols are changed by − 0.02, + 0.06 and + 0.02 K on average in the region (100°–130° E, 20°–50° N) between the strongest and weakest EASM index years. Similarly, the enhancement in EASM circulation induced by EM (IM) aerosols is weaker (stronger) in the strongest EASM index years than in the weakest years (not shown here). Relevant studies indicated that the total aerosol combined effects on the EASM were more significant in stronger EASM years (Li et al. 2016a, b) and that the interaction between the BC warming effect and EASM was more intense during weaker monsoon index years (Zhuang et al. 2018a). The results here show much more complicated influences of the direct aerosol effect because of the variations in the aerosol mixtures.

Fig. 13
figure 13

Changes in air temperature (shaded; K) in the lower troposphere averaged from the surface to 850 hPa from the externally (left, EM), internally (middle, IM) and partially internally (right, PIM) mixed aerosol direct effects over East Asia during the weakest East Asian summer monsoon index years (top panel) and between the strongest and weakest monsoon index years (middle panel), as well as the differences of precipitation responses (shaded; mm/d at bottom panel) from the externally (left, EM), internally (middle, IM) and partially internally (right, PIM) mixed aerosol direct effects over East Asia between the strongest and weakest monsoon index years

Full size image

Different definitions of East Asian summer monsoon index might result in different interannual variations of the monsoon. Here, another widely used EASM index proposed by Zhang et al. (2003) is further discussed. This index is defined as a seasonal difference of anomalous zonal wind between the regions in (100°–150° E, 10°–20° N) and (100°–150° E, 25°–35° N) at 850 hPa. Based on their definition, 1997, 2002 and 2004 are classified as strong monsoon index years and 1998, 2003 and 2010 are classified as weak monsoon index years, which is consistent with Li et al. (2016a, b). Comparisons indicate that the differences of atmospheric circulations, aerosol loadings and precipitations between in stronger and weaker index years are very similar to the ones shown in Figs. 11 and 12, implying that the differences of the aerosol climate and environmental effects between the stronger and weaker EASM index years of Zhang et al. (2003) might be similar to the differences analyzed in this section. Nevertheless, further influences with more EASM index choices are needed to be investigated to fully address this issue in the future.

4 Conclusions

Due to the lack of observations, aerosol mixing issues are still not well addressed, which may result in certain uncertainties in estimating aerosol radiative forcing, regional climate effects, and environmental effects (Forster et al. 2007). To better understand the interactions between air pollution and regional climate change. The direct effects of the different mixed anthropogenic aerosols on the East Asian summer monsoon (EASM) are further investigated in this study by employing a well-developed regional climate and chemistry modeling system.

The results indicate that high levels of anthropogenic aerosols are mainly distributed in southwest and north China in summer. The regional mean total aerosol surface concentration in the East Asian region is about 14.13 μg/m3 in summer, in which sulfate accounts for the most. The surface BC loading in this region may account for approximately 13% of the total aerosols, implying the importance of aerosol mixtures in assessing the aerosol climate effect. The aerosol mixing states can result in a variation of 0.27 for the regional mean aerosol single scattering albedo (SSA), which subsequently induces large variations in the aerosol direct radiative forcing (DRF). The regional mean aerosol instantaneous DRFs at the TOA can be changed from − 0.95 W/m2 to + 1.50 W/m2 in summer in East Asia when the aerosols are from the external mixing state (EM) to the internal mixing state (IM). Regional climate feedback may weaken the influence of aerosol mixings on aerosol radiative forcing. Previous investigations have stated that partially internally mixed (PIM) aerosols are closer to real conditions. The aerosol SSA and instantaneous and effective DRFs at the TOA are 0.95, − 0.12 W/m2 and − 5.16 W/m2, respectively, in summer.

Due to the absorption of BC, the effective DRFs are all positive in the atmosphere. The more aerosols are mixed internally, the stronger the positive DRFs, and a larger shortwave heating rate (SHR) increase will be found. The aerosol mixing states may cause a variation of 0.08 K/day for the regional mean lower atmospheric SHR in summer. Investigations here indicate that the direct effect of anthropogenic aerosols on the East Asian climate is more similar to that of BC in summer, even though BC only accounts for a small proportion of the total aerosols. Low layer air temperatures are increased considerably, hence resulting in decreases in cloud cover below 850 hPa and precipitation in most areas of East Asia. In contrast, the East Asian summer monsoon circulation is enhanced, leading to a local increase in rainfall and a smaller decrease in cloud amount in South China. As a result of the EASM anomalies, all the aerosol components are increased substantially. Similarly, the direct effect of anthropogenic aerosols on the EASM and regional air qualities will become increasingly significant with an increase in the internal mixing proportion of aerosols. The regional responses of the lower atmospheric air temperature, cloud amount, wind speed, precipitation, and surface aerosol loadings in East Asia can vary from 0.26 to 0.34 K, 0.46 to 0.53%, 0.11 to 0.30 m/s, − 0.05 to − 0.01 mm/day, and 11.04 to 12.03%, respectively, because of different aerosol mixing states. Where the PIM aerosols result in a change of + 0.28 K, − 0.04 mm/day and + 1.56 μg/m3for the lower atmospheric temperature, precipitation and total aerosol surface concentrations, respectively, in East Asia in summer.

Due to considerable internal-annual variation in the EASM, the surface concentrations have a more than 25% difference for secondary inorganic aerosols and a more than 8.7% difference for primary carbonaceous aerosols in the East Asia region in summer between the strongest and weakest monsoon index years. As a result, the interaction between the anthropogenic aerosol direct effect and EASM will be different in different monsoon index years. The investigations further indicate that the responses of the EASM to aerosols between the strongest and weakest monsoon index years are also affected by the aerosol mixing states. Overall, the enhancement of the EASM circulation induced by the EM aerosols is weaker in the strongest monsoon index years, which is opposite to that induced by the IM or PIM aerosols.

Although the direct effect of anthropogenic aerosols, as well as its uncertainties, is investigated in this study, limitations still exist. For example: 1. Similar to other studies (Lesins et al. 2002; Riemer et al. 2003; Stier et al. 2006; Matsui et al. 2013; Yu et al. 2013; Saleh et al. 2015; Zhou et al. 2018; Curci et al. 2019; etc.), not all of the aerosol species have been taken into account in this study when investigating the mixed aerosol radiative effects. Although only anthropogenic aerosols are taken into account here, the findings are still representative to a certain degree. Previous studies have indicated that the AODs of dust, ammonium and secondary organic carbon aerosols, which were not included here, were quite small in East Asia in summer (Sun et al. 2012; Xu and Penner 2012; Park et al. 2014; and Yin et al. 2015). For these reasons, the simulated AOD in East Asia in summer matches well with MODIS observation in this study, based on the well simulations of the model to sulfate, nitrate, primary carbonaceous aerosol loadings. Therefore, the major aerosol characteristics in East Asia in summer have been captured in our simulations. Additionally, sensitive tests from Curci et al. (2019) indicated that changes in secondary inorganic (scattering) aerosol loadings would yields a negligible change in terms of calculated absorption properties (SSA) when aerosols were in partial internal mixing situation and the bias of scattering aerosol and black carbon aerosol mass might be unlikely to alter the main conclusions regarding calculated absorption properties. 2. As it has been depicted in many studies, to fully explore the impacts of mixed aerosol is a complex task (Zhou et al. 2018). Thus, the internal mixing ratio of the aerosols is homogeneous over the entire region owing to a lack of observations. 3. The interaction between direct effects of mixing aerosols and the East Asian monsoon in winter is unclear. 4. The indirect aerosol effects, which are also important to regional climate changes, are excluded here. These issues should be addressed in further studies to better understand the effect of aerosol mixing states on regional climate changes and air pollution variations.