Atmospheric aerosols can exert an influence on meteorology and air quality through aerosol–radiation interaction (ARI) and aerosol–cloud interaction (ACI), and this two-way feedback has been studied by applying two-way coupled meteorology and air quality models. As one of the regions with the highest aerosol loading in the world, Asia has attracted many researchers to investigate the aerosol effects with several two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) over the last decade. This paper attempts to offer a bibliographic analysis regarding the current status of applications of two-way coupled models in Asia, related research focuses, model performances, and the effects of ARI and/or ACI on meteorology and air quality. There were a total of 160 peer-reviewed articles published between 2010 and 2019 in Asia meeting the inclusion criteria, with more than 79 % of papers involving the WRF-Chem model. The number of relevant publications has an upward trend annually, and East Asia, India, and China, as well as the North China Plain are the most studied areas. The effects of ARI and both ARI and ACI induced by natural aerosols (particularly mineral dust) and anthropogenic aerosols (bulk aerosols, different chemical compositions, and aerosols from different sources) are widely investigated in Asia. Through the meta-analysis of surface meteorological and air quality variables simulated by two-way coupled models, the model performance affected by aerosol feedbacks depends on different variables, simulation time lengths, selection of two-way coupled models, and study areas. Future research perspectives with respect to the development, improvement, application, and evaluation of two-way coupled meteorology and air quality models are proposed.
Atmospheric pollutants can affect local weather and global climate via many mechanisms, as extensively summarized in the Intergovernmental Panel on Climate Change (IPCC) reports (IPCC, 2007, 2013, 2021), and also exhibit impacts on human health and ecosystems (Lelieveld et al., 2015; Wu and Zhang, 2018). Atmospheric pollutants can modify the radiation energy balance, thus influencing meteorological conditions (Gray et al., 2010; Yiğit et al., 2016). Compared to other climate agents, short-lived and localized aerosols could induce changes in meteorology and climate through aerosol–radiation interaction (ARI; Tremback et al., 1986; Satheesh and Moorthy, 2005) and aerosol–cloud interaction (ACI; Martin and Leight, 1949; Lohmann and Feichter, 2005) or both (Sud and Walker, 1990; Haywood and Boucher, 2000). ARI (previously known as the direct effect and semi-direct effect) is based on scattering and absorbing solar radiation by aerosols as well as cloud dissipation by heating (McCormick and Ludwig, 1967; Ackerman et al., 2000; Koch and Del Genio, 2010; Wilcox, 2012). ACI (known as the indirect effect) is concerned with aerosols altering albedo and lifetime of clouds (Twomey, 1977; Albrecht, 1989; Lohmann and Feichter, 2005). As our knowledge base of aerosol–radiation–cloud interactions that involve extremely complex physical and chemical processes has been expanding, accurately assessing the effects of these interactions still remains a big challenge (Rosenfeld et al., 2008, 2019; Fan et al., 2016; Kuniyal and Guleria, 2019).
The interactions between air pollutants and meteorology can be investigated by observational analyses and/or air quality models. So far, many observational studies using measurement data from a variety of sources have been conducted to analyze these interactions (Wendisch et al., 2002; Bellouin et al., 2008; Groß et al., 2013; Rosenfeld et al., 2019). Yu et al. (2006) reviewed research work that adopted satellite and ground-based measurements to estimate the ARI-induced changes in radiative forcing and the associated uncertainties in the analysis. Yoon et al. (2019) analyzed the effects of aerosols on the radiative forcing based on the Aerosol Robotic Network observations and demonstrated that these effects depend on aerosol types. On the other hand, since the uncertainties in ARI estimations have been associated with ACI (Kuniyal and Guleria, 2019), simultaneous assessments of both ARI and ACI effects are needed and have gradually been conducted via satellite observations (Sekiguchi et al., 2003; Quaas et al., 2008; Illingworth et al., 2015; Kant et al., 2019). In the early stages, observational studies of ACI effects were based on several cloud parameters mainly derived from surface-based microwave radiometer (Kim et al., 2003; Liu et al., 2003) and cloud radar (Feingold et al., 2003; Penner et al., 2004). Later on, with the further development of satellite observation technology and enhanced spatial resolution of satellite measurement compared against traditional ground observations, satellite-retrieved cloud parameters (effective cloud droplet radius, liquid water path – LWP, and cloud cover) were utilized to identify the ACI effects studies on a cloud scale (Goren and Rosenfeld, 2014; Rosenfeld et al., 2014). Moreover, in order to clarify whether aerosols affect precipitation positively or negatively, the effects of ACI on cloud properties and precipitation were widely investigated but with various answers (Andreae and Rosenfeld, 2008; Rosenfeld et al., 2014; Casazza et al., 2018; Fan et al., 2018). Analyses of satellite and/or ground observations revealed that increased aerosols could suppress (enhance) precipitation in drier (wetter) environments (Rosenfeld, 2000; Rosenfeld et al., 2008; Z. Li et al., 2011; Donat et al., 2016). Most recently, Rosenfeld et al. (2019) further used satellite-derived cloud information (droplet concentration and updraft velocity at cloud base, LWP at cloud cores, cloud geometrical thickness, and cloud fraction) to single out ACI under a certain meteorological condition and found that the cloudiness change caused by aerosol in marine low-level clouds was much greater than previous analyses (Sato and Suzuki, 2019). Despite the fact that the aforementioned studies significantly improved our understanding of aerosol effects, many limitations still exist, such as low temporal resolution of satellite data, low spatial resolution of ground monitoring sites, and lack of vertical distribution information on aerosol and cloud (Yu et al., 2006; Rosenfeld et al., 2014; Sato and Suzuki, 2019).
Numerical models can also be used to study the interactions between air
pollutants and meteorology. Air quality models simulate physical and
chemical processes in the atmosphere (ATM) and are classified as offline and
online models (El-Harbawi, 2013). Offline models (also known as traditional
air quality models) require outputs from meteorological models to
subsequently drive chemical models (Seaman, 2000; Byun and Schere, 2006;
Ramboll Environment and Health, 2008). Compared to online models, offline models are usually
computationally efficient but incapable of capturing two-way feedbacks
between chemistry and meteorology (North et al., 2014). Online models or
coupled models are designed and developed to consider the two-way feedbacks
and have attempted to accurately simulate both meteorology and air quality (Grell
et al., 2005; Wong et al., 2012; Briant et al., 2017). Two-way coupled
models can be generally categorized as integrated and access models based on
whether they use a coupler to exchange variables between meteorological and
chemical modules (Baklanov et al., 2014). As Zhang (2008) pointed out,
Jacobson (1994, 1997a) and Jacobson et al. (1996a) pioneered the development
of a fully coupled model named the Gas, Aerosol, Transport, Radiation, General
Circulation, Mesoscale, and Ocean Model (GATOR-GCMOM) in order to
investigate all the processes related to ARI and ACI. Currently, there are
three representative two-way coupled meteorology and air quality models,
namely the Weather Research and Forecasting-Chemistry (WRF-Chem) (Grell et
al., 2005), WRF coupled with Community Multiscale Air Quality (CMAQ) (Wong
et al., 2012), and WRF coupled with a multi-scale chemistry-transport model
for atmospheric composition analysis and forecast (WRF-CHIMERE) (Briant et
al., 2017). WRF-Chem is an integrated model that includes various
chemical modules in the meteorological model (i.e., WRF) without using a
coupler. For the remaining two models, which are access models,
WRF-CMAQ uses a subroutine called
Compared to the US and Europe, Asia has been suffering from more severe air pollution
in the past 3 decades (Bollasina et al., 2011; Rohde and Muller, 2015;
Gurjar et al., 2016) due to the rapid industrialization, urbanization, and
population growth together with unfavorable meteorological conditions (Jeong
and Park, 2017; Li et al., 2017a; Lelieveld et al., 2018). The
interactions between atmospheric pollution and meteorology in Asia, which
have received a lot of attention from the scientific community, are investigated
using extensive observations and a certain number of numerical simulations
(Wang et al., 2010; Li et al., 2016; Nguyen et al., 2019a). Based on
airborne, ground-based, and satellite-based observations, multiple important
experiments have been carried out to analyze properties of radiation, cloud,
and aerosols in Asia, as briefly reviewed by Lin et al. (2014b). Recent
observational studies confirmed that increasing aerosol loadings play
important roles in the radiation budget (Eck et al., 2018; Benas et al., 2020),
cloud properties (Dahutia et al., 2019; Yang et al., 2019), and precipitation
intensity along with vertical distributions of precipitation types (Guo et
al., 2014, 2018). According to previous observational studies in Southeast
Asia (SEA), Tsay et al. (2013) and Lin et al. (2014b) comprehensively
summarized the spatiotemporal characteristics of biomass burning (BB)
aerosols and clouds as well as their interactions. Li et al. (2016) analyzed
how ARI or ACI influenced climate and meteorology in Asia utilizing observations
and climate models. With regard to the impacts of aerosols on cloud,
precipitation, and climate in East Asia (EA), a detailed review of
observations and modeling simulations has also been presented by Z. Li et al. (2019). Since the 2000s, substantial progress has been made in
climate–air pollution interactions in Asia based on regional climate model
simulations, which have been summarized by Li et al. (2016). Moreover,
starting from the year 2010, with the development and availability of two-way
coupled meteorology and air quality models, more and more modeling studies
have been conducted to explore the ARI and/or ACI effects in Asia (H. Wang et
al., 2010; J. Wang et al., 2014; Sekiguchi et al., 2018; Nguyen et al., 2019a). In recent
studies, a series of WRF-Chem and WRF-CMAQ simulations were performed to
assess the consequences of ARI for radiative forcing, planetary boundary
layer height (PBLH), precipitation, and fine particulate matter (PM
This paper is constructed as follows: Sect. 2 describes the methodology for literature searching, paper inclusion, and analysis; Sect. 3 summarizes the basic information about publications as well as developments and applications of coupled models in Asia, and Sect. 4 provides the recent overviews of their research points. Sections 5 to 6 present a systematic review and meta-analysis of the effects of aerosol feedbacks on model performance, meteorology, and air quality in Asia. The summary and perspective are provided in Sect. 7.
Since 2010, in Asia, regional studies of aerosol effects on meteorology and air quality based on coupled models have been increasing gradually; therefore, in this study we performed a systematic search of the literature to identify relevant studies from 1 January 2010 to 31 December 2019. In order to find all the relevant papers in English, Chinese, Japanese, and Korean, we deployed serval science-based search engines, including Google Scholar, the Web of Science, the China National Knowledge Infrastructure, the Japan Information Platform for S&T Innovation, and the Korean Studies Information Service System. The different keywords and their combinations for paper searching are as follows: (1) model-related keywords including “coupled model”, “two-way”, “WRF”, “NU-WRF”, “WRF-Chem”, “CMAQ”, “WRF-CMAQ”, “CAMx”, “CHIMERE”, “WRF-CHIMERE” and “GATOR-GCMOM”; (2) effect-related keywords including “aerosol radiation interaction”, “ARI”, “aerosol cloud interaction”, “ACI”, “aerosol effect”, and “aerosol feedback”; (3) air-pollution-related keywords including “air quality”, “aerosol”, “PM2.5”, “O3”, “CO”, “SO2”, “NO2”, “dust”, “BC”, “black carbon”, “blown carbon”, “carbonaceous”, and “primary pollutants”; (4) meteorology-related keywords including “meteorology”, “radiation”, “wind”, “temperature”, “specific humidity”, “relative humidity”, “planetary boundary layer”, “cloud”, and “precipitation”; (5) region-related keywords including “Asia”, “East Asia”, “Northeast Asia”, “South Asia”, “Southeast Asia”, “Far East”, “China”, “India”, “Japan”, “Korea”, “Singapore”, “Thailand”, “Malaysia”, “Nepal”, “North China Plain”, “Yangtze River Delta”, “Pearl River Delta”, “middle reaches of the Yangtze River”, “Sichuan Basin”, “Guanzhong Plain”, “Northeast China”, “Northwest China” “East China”, “Tibet Plateau”, “Taiwan”, “northern India”, “southern India”, “Gangetic Basin”, and “Kathmandu Valley”.
After applying the search engines and the keyword combinations mentioned above, we found 946 relevant papers. In order to identify which papers should be included or excluded in this paper, the following criteria were applied: (1) duplicate literature was deleted; (2) studies using coupled models in Asia with aerosol feedbacks turned on were included, and observational studies of aerosol effects were excluded; (3) publications involving coupled climate models were excluded. According to these criteria, not only regional studies, but also studies using the coupled models at global or hemispheric scales involving Asia or its subregions were included. Then, we carefully examined all the included papers and further checked the listed references in each paper to make sure that no related paper was neglected. A flowchart that illustrates the detailed procedures applied for article identification is presented in Fig. A1 (note: although the deadline for literature searching is 2019, any literature published in 2020 is also included). There were a total of 160 publications included in our study.
To summarize the current status of coupled models applied in Asia and quantitatively analyze the effects of aerosol feedbacks on model performance as well as meteorology and air quality, we carried out a series of analyses based on data extracted from the selected papers. We firstly compiled the publication information from the included papers as well as the information regarding model name, simulated time period, study region, simulation design, and aerosol effects. Secondly, we summarized the important findings of two-way coupled model applications in Asia according to different aerosol sources and components to clearly determine the major research focuses in past studies. Finally, we gathered all the simulated results of meteorological and air quality variables with and without aerosol effects and their statistical indices (SIs). For questionable results, quality assurance was conducted after personal communications with the original authors to decide whether they were deleted and/or corrected. All the extracted publication and statistical information was exported into an Excel file, which is provided in Table S1. Moreover, we performed quantitative analyses of the effects of aerosol feedbacks through the following steps. (1) We discussed whether meteorological and air quality variables were overestimated or underestimated based on their SIs. Then, variations of the SIs of these variables were further analyzed in detail with and without turning on ARI and/or ACI in two-way coupled models. (2) We investigated the SIs of simulation results at different simulation time lengths and spatial resolutions in coupled models. (3) More detailed inter-model comparisons of model performance based on the compiled SIs among different coupled models are conducted. (4) Differences in simulation results with and without aerosol feedbacks were grouped by study regions and timescales (yearly, seasonal, monthly, daily, and hourly). Toward a better understanding of the complicated interactions between air quality and meteorology in Asia, the results sections in this paper are organized following the above analysis methods (1)–(3) and presented in Sect. 5, and the results following method (4) are presented in Sect. 6. In addition, Excel and Python were used to conduct data processing and plotting in this study.
A total of 160 articles were selected according to the inclusion criteria, and their basic information was compiled in Table 1. In Asia, five two-way coupled models are applied to study the ARI and ACI effects. These include GATOR-GCMOM, two commonly used models, i.e., WRF-Chem and WRF-CMAQ, and two locally developed models, i.e., the global–regional assimilation and prediction system coupled with the Chinese Unified Atmospheric Chemistry Environment forecasting system (GRAPES-CUACE) and WRF coupled with the nested air quality prediction modeling system (WRF-NAQPMS). A total of 127 out of 160 papers involved the applications of WRF-Chem in Asia since its two-way coupled version was publicly available in 2006 (Fast et al., 2006). WRF-CMAQ was applied in only 16 studies due to its later initial release in 2012 (Wong et al., 2012). GRAPES-CUACE was developed by the China Meteorological Administration and introduced in detail in Zhou et al. (2008, 2012, 2016), then firstly utilized in Wang et al. (2010) to estimate impacts of aerosol feedbacks on meteorology and the dust cycle in EA. The coupled version of WRF-NAQPMS was developed by the Institute of Atmospheric Physics, Chinese Academy of Sciences, and improved the prediction accuracy of haze pollution in the North China Plain (NCP) (Z. Wang et al., 2014). Note that GRAPES-CUACE and WRF-NAQPMS were only applied in China. There were only three published papers about the applications of GATOR-GCMOM in northeastern Asia (NEA), NCP, and India. In the included papers, 93, 33, and 31 studies targeted various areas in China, EA, and India, respectively. There were 79 papers regarding effects of ARI (7 health), 63 for both ARI and ACI (1 health), and 18 for ACI. ACI studies were much fewer than ARI-related ones, which indicated that ACI-related studies need to be paid more attention in the future. Considering that the choices of cloud microphysics and radiation schemes can affect coupled models' results (Baró et al., 2015; Jimenez et al., 2016), the schemes used in the selected studies are also summarized in Table 1. This table presents a concise overview of coupled models' applications in Asia with the purpose of providing basic information regarding models, study periods and areas, aerosol effects, scheme selections, and references. More complete information is summarized in Table S1 including model version, horizontal resolution, vertical layer, aerosol- and gas-phase chemical mechanisms, photolysis rate, PBL, land surface, surface layer, cumulus, urban canopy schemes, meteorological initial and boundary conditions (ICs and BCs), chemical ICs and BCs, spin-up time, and anthropogenic and natural emissions.
It should be noted that in Table 1 there are four model intercomparison
studies that aimed at evaluating model performance, identifying error
sources and uncertainties, and providing optimal model setups. By comparing
simulations from two coupled models (WRF-Chem and Spectral
Radiation-Transport Model for Aerosol Species) (Takemura et al., 2003) in
India (Govardhan et al., 2016), it was found that the spatial distributions
of various aerosol species (black carbon –BC, mineral dust, and sea salt)
were similar with the two models. Based on the intercomparisons of WRF-Chem
simulations in different areas, Yang et al. (2017) revealed that aerosol
feedbacks could enhance PM
Basic information on coupled model applications in Asia during 2010–2019.
Continued.
Continued.
Continued.
Continued.
Continued.
To gain an overall understanding of applications of coupled models in Asia, the spatial distributions of study areas from the selected literature and the temporal variations of the annual publication numbers were extracted from Table 1 and summarized. Figure 1 illustrates the spatial distributions of study regions as well as the number of papers involving coupled models in Asia (Fig. 1a) and China (Fig. 1b). In this figure, the color and number in the pie charts represent individual (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) or multiple coupled models and the quantity of corresponding articles, respectively. At subregional scales, most studies targeted EA where high anthropogenic aerosol loading occurred in recent decades, mainly using WRF-Chem and WRF-CMAQ (Fig. 1a). For other subregions, such as NEA, SEA, central Asia (CA), and western Asia (WA), there were rather limited research activities taking into account aerosol feedbacks with two-way coupled models. National-scale applications of two-way coupled models targeted mostly modeling domains covering India and China, but much less work has been carried out in other countries, such as Japan and Korea, where air pollution levels are much lower. With respect to various areas in China (Fig. 1b), the research activities concentrated mostly in NCP and secondly in eastern China (EC), then in the Yangtze River Delta (YRD) and Pearl River Delta (PRD) areas. WRF-Chem was the most popular model applied in all areas, but there were a few applications of GPRAPES-CUACE and WRF-NAQPMS in EC and NCP.
Figure 2 depicts the temporal variations of research activities with two-way coupled models in Asia over the period of 2010 to 2019. The total number of papers related to two-way coupled models had an obvious upward trend in the past decade. Prior to 2014, applications of two-way coupled models in Asia were scarce, with about one to six publications per year. A noticeable increase in research activities emerged starting from 2014, and the growth was rapid from 2014 to 2016 at a rate of seven to nine more papers per year, especially in China. It could be related to the Action Plan on Prevention and Control of Atmospheric Pollution (2013–2017) implemented by the Chinese government. The growth was rather flat during 2016–2018 before reaching a peak of 31 articles in 2019. In addition, the pie charts in Fig. 2 indicate that modeling activities had been picking up with a diversified pattern in the study domain from 2010 to 2019. The modeling domains extended from EA to China and India and then several subregions in Asia and various areas in China. For EA and India, investigations of aerosol feedbacks based on two-way coupled models rose from one to two papers per year during 2010–2013 to four to eight during 2014–2019. Since 2014, most model simulations were carried out with a focus on areas with severe air pollution in China, especially the NCP area with five to seven publications per year.
The spatial distributions of study domains as well as the two-way
coupled modeling publication numbers in different subregions or countries of
Asia
The temporal variations of study activities adopting two-way coupled models in Asia during 2010–2019. (EA: East Asia, NEA: northeastern Asia, SEA: Southeast Asia, EC: eastern China, NCP: North China Plain, YRD: Yangtze River Delta, SEC: southeastern China, NWC: northwestern China, TP: Tibetan Plateau, MRYR: middle reaches of the Yangtze River, SWC: southwestern China; PRD: Pearl River Delta).
The physiochemical processes involved with ARI and ACI are sophisticated in actual conditions of the atmospheric environment, but their representations in two-way coupled models can be rather different. Also, simulation results depend on how these models are configured and set up. Therefore, the treatments of aerosol and cloud microphysics, aerosol–radiation–cloud interactions in WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM applied in Asia, and the various aspects of how the modeling studies are set up in the selected papers are summarized in Tables 2–5, respectively, and outlined in this section.
Aerosol microphysics processes consist of particle nucleation, coagulation,
condensation and evaporation, gas–particle mass transfer, inorganic aerosol
thermodynamic equilibrium, aqueous chemistry, and formation of secondary
organic aerosol (SOA). Their representations in a variety of aerosol
mechanisms offered in the five two-way coupled models applied in Asia and
relevant references are compiled in Table 2. Note that the GOCART scheme in
WRF-Chem is based on a bulk aerosol mechanism that is not able to consider
the details of these microphysics processes. The binary homogeneous
nucleation schemes with and without hydration developed by different authors are
applied in the five coupled models for simulating new particle formation,
and GATOR-GCMOM also adopts the ternary nucleation parameterization scheme
for H
Treatments of aerosol microphysics processes in two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) applied in Asia.
Continued.
In addition to aerosol microphysics processes, the cloud properties included in cloud microphysics schemes and the treatment of aerosol–cloud processes in the five two-way coupled models are different in terms of hydrometeor classes, cloud droplet size distribution, aerosol water uptake, in- and below-cloud scavenging, hydrometeor–aerosol coagulations, and sedimentation of aerosols and cloud droplets (Table 3). Among the microphysics schemes implemented in the five coupled models, mass concentrations of different hydrometeors (including cloud water, rain, ice, snow, or graupel) are included, but their number concentrations are only considered if the cloud microphysics schemes are two-moment or three-moment. The single modal approach with either lognormal or gamma distribution and the sectional approach with discrete size distributions for cloud droplets are applied in different microphysics schemes. Based on the Mie theory, WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM calculate cloud radiative properties (including the extinction, scattering, and absorption coefficients, single-scattering albedo, and asymmetry factor of liquid and ice clouds) in their radiation schemes (e.g., RRTMG, GODDARD, GATOR2012). In the atmosphere, the hygroscopic growth of aerosols due to water uptake is parameterized based on the Köhler or Zdanovskii–Stokes–Robinson theory, and the hysteresis effects depending on the deliquescence and crystallization RH are taken into account in the five coupled models. The removal processes of aerosol particles include wet removal and sedimentation. Aerosol particles in accumulation and coarse modes can act as cloud condensation nuclei (CCN) or ice nuclei (IN) via activations in cloud, which can further develop to different types of hydrometeors (cloud water, rain, ice, snow, and graupel) and then gradually form precipitation. These processes are called in-cloud scavenging or rainout. The aerosol particles below cloud base also can be coagulated with the falling hydrometeors, which is known as below-cloud scavenging or washout. Representations of both in- and below-cloud scavenging processes are based on the scavenging rate approach in aerosol mechanisms of WRF-Chem, WRF-CMAQ, GRAPES-CUACE, and WRF-NAQPMS but not GATOR-GCMOM. Size-resolved sedimentation of aerosols is computed from one model layer to layers below down to the surface layer using setting velocity in most coupled models, and the MOSAIC aerosol mechanism in WRF-Chem only considers the sedimentation in the lowest model level (Marelle et al., 2017).
Compilation of cloud properties and aerosol–cloud processes in two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) applied in Asia.
Continued.
Table 4 further lists various aspects with regards to how ARI and ACI are calculated in the five two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) applied in Asia. Note that the information in this table was extracted from the latest released version of WRF-Chem (version 4.3.3) and WRF-CMAQ (based on WRF v4.3 and CMAQ v5.3.3) as well as relevant references for GRAPES-CUACE (H. Wang et al., 2015), WRF-NAQPMS (Z. Wang et al., 2014), and GATOR-GCMOM (Jacobson, 2012a). These models all use the Mie theory to compute ARI effects but differ in representations of aerosol optical properties and radiation schemes. To simplify the calculation, aerosol species simulated by the chemistry module and/or model are put into different groups (Table 4), and the refractive indices of these groups are directly from the Optical Properties of Aerosols and Clouds (OPAC) database (Hess et al., 1998) in WRF-Chem and WRF-CMAQ (Table B6 in Appendix B). In WRF-Chem, the aerosol optical properties (AOD, extinction, scattering, and absorption coefficients, single-scattering albedo, and asymmetry factor) are calculated in terms of four spectral intervals (listed in Table B6 in Appendix B) and then interpolated and/or extrapolated to 11 (14) SW intervals defined in the GODDARD (RRTMG) scheme. For SW and LW radiation in both WRF-CMAQ and WRF-Chem, these optical parameters are computed at each of the corresponding spectral intervals in the RRTMG scheme. The aerosol optical property for LW radiation is considered only at five thermal windows (listed in Table B6) in WRF-CMAQ. No detailed information regarding how aerosol optical properties and relevant parameters are calculated in GRAPES-CUACE and WRF-NAQPMS can be found from the relevant references.
With respect to ACI effects, the simulated aerosol characteristics (such as mass, size distribution and species) are utilized for the calculation of cloud droplet activation and aerosol resuspension based on the Köhler theory (Abdul-Razzak and Ghan, 2002) in several (one) microphysics schemes (scheme) in WRF-Chem (GRAPES-CUACE). GATOR-GCMOM is the first two-way coupled model adding IN activation processes including heterogeneous and homogeneous freezing (Jacobson, 2003). None of the other four two-way coupled models consider the IN formation processes (including immersion freezing, deposition freezing, contact freezing, and condensation freezing), but they have been included in some specific versions of WRF-Chem (Keita et al., 2020; Lee et al., 2020), which are not yet in the latest release version 4.3.3 of WRF-Chem.
Summary of relevant information regarding calculations of aerosol–radiation interaction (ARI) and aerosol–cloud interaction (ACI) in two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) applied in Asia.
How accurately ARI and ACI are simulated also relies on the representation of
aerosol composition and size distribution in two-way coupled models. Table 5
presents the treatments of aerosol compositions and size distributions in
the five two-way coupled models applied in Asia. As shown in Tables 4 and 5,
GATOR-GCMOM considered more detailed aerosol species groups with as many as 42
kinds, and other coupled models considered different numbers of species groups (such
as six, five, seven, and eight aerosol species groups in WRF-Chem, CMAQ, NAQPMS, and CUACE,
respectively). Three typical representation approaches of size distribution
(bulk, modal, and sectional methods) are adopted by the five two-way coupled
models, and WRF-Chem offers all three approaches, but other models only
support one specific option. The Global Ozone Chemistry Aerosol Radiation
and Transport (GOCART) model (Ginoux et al., 2001) in WRF-Chem is the only
one that is based on a combination of bulk (for water, BC, OC, and sulfate
aerosols) and sectional (for dust and sea salt aerosols) approaches. The
widely used modal and sectional approaches in five coupled models and their
detailed numerical settings of aerosol size distribution (namely, geometric
diameter and standard deviation for the modal approach and bin ranges for
the sectional method) are listed in Table 5. Regarding the modal method, same
parameter values for Aitken and accumulation modes as well as geometric diameters
for the coarse mode in the latest version of WRF-Chem (v4.3.3) and the older version
of WRF-CMAQ (before v5.2) are set as default, except the standard deviations
for the coarse mode, which are slightly different. In the official version of WRF-CMAQ
released after v5.2, there are some modifications to the default setting of
geometric diameters in Aitken, accumulation, and coarse modes from 0.01 to
0.015, 0.07 to 0.08, and 1.0 to 0.6
Summary of numerical representations of aerosol size distribution and composition in two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) applied in Asia.
Continued.
Not only the choice of methodologies for ARI and ACI calculations can impact simulation results, but also the various aspects regarding the setup of modeling studies by applying two-way coupled models. The extra and/or auxiliary information about model configuration, including horizontal and vertical resolutions, aerosol- and gas-phase chemical mechanisms, PBL schemes, meteorological and chemical ICs and BCs, and anthropogenic and natural emissions, were extracted from the 160 papers and are presented in Table S4 of the Supplement, which is organized in the same order as Table 1.
For two-way coupled model applications in Asia, horizontal resolutions were
set from a few to several hundred kilometers, sometimes with nests, and
vertical resolutions were from 15 to about 50–70 levels, with only one study
performed at 100 levels for studying a fog case (Z. Wang et al., 2019). K. Wang
et al. (2018) evaluated the impacts of horizontal resolutions on simulation
results and found that surface meteorological variables were better modeled
at finer resolution, but there were no significant improvements of ACI-related
meteorological variables and certain chemical species between different grid
resolutions. By applying a single column model and then WRF-Chem with
ARI, Z. Wang et al. (2019) revealed that better representation of PBL
structure and relevant variables with finer vertical resolution from the
surface to the PBL top could reduce model biases noticeably, but balancing
between vertical resolution and computational resources was important as
well. Among the 160 applications of two-way coupled models in Asia, the
frequently used aerosol module and gas-phase chemistry mechanism in WRF-CMAQ
(WRF-Chem) were AERO6 (MOSAIC and MADE/SORGAM) and CB05 (CBMZ and RADM2),
respectively. For PBL schemes, most studies selected YSU in WRF-Chem and
ACM2 in WRF-CMAQ. Regarding meteorological ICs and BCs, the FNL data were
the first choice, and outputs from the Model for Ozone and Related Chemical
Tracer (MOZART) were used to generate chemical ICs and BCs by most
researchers. Georgiou et al. (2018) also revealed that boundary conditions
of dust and O
Due to the fact that dust storm events frequently occurred over Asia during 2000–2010, the research community has focused on dust transportation and associated climatic effects (Gong et al., 2003b; Zhang et al., 2003a, b; Yasunari and Yamazaki, 2009; Lee et al., 2010; Choobari et al., 2014). Also, the detailed processes and physiochemical mechanisms of dust storms have been well understood and reviewed in detail (Shao and Dong, 2006; Uno et al., 2006; Huang et al., 2014; S. Chen et al., 2017b). To probe the radiative feedbacks of dust aerosols in Asia, Wang et al. (2010, 2013) initiated modeling studies by a two-way coupled model, i.e., the GRAPES-CUACE model, to simulate direct radiative forcing (DRF) of dust and revealed that the feedback effects of dust aerosols could lead to decreasing surface wind speeds and then suppress dust emissions. Further modeling simulations by the same model (Wang and Niu, 2013) indicated that considering dust radiative effects did not substantially improve the model performance of the air temperature at 2 m above the surface (T2), even when assimilating data from in situ and satellite observations into the model. Subsequently, several similar studies based on another two-way coupled model (WRF-Chem with the GOCART scheme) were conducted to investigate dust radiative forcing (including shortwave radiative forcing – SWRF – and longwave radiative forcing – LWRF) and ARI effects of dust on meteorological variables (PBLH, T2 and WS10) in different regions of Asia (Kumar et al., 2014; Chen et al., 2014; Jin et al., 2015, 2016b; L. Liu et al., 2016; Bran et al., 2018; Su and Fung, 2018a, b; Zhou et al., 2018). These studies demonstrated that dust aerosols could induce negative radiative forcing (cooling effect) at the top of the atmosphere (TOA) as well as the surface (including both Earth's and sea surfaces) and positive radiative forcing (warming effect) in the ATM (Wang et al., 2013; Chen et al., 2014; Kumar et al., 2014; M. M. Li et al., 2017a; Bran et al., 2018; Li and Sokolik, 2018; Su and Fung, 2018b). More thorough analyses of the radiative effects of dust in Asia (Wang et al., 2013; Li and Sokolik, 2018) pointed out that dust aerosols played opposite roles in the shortwave and longwave bands so that the dust SWRF at TOA and the surface (cooling effects) as well as in the ATM (warming effects) was offset partially by the dust LWRF (warming effects at TOA and the surface but cooling effects in the ATM). It was noteworthy that adding a more detailed mineralogical composition to the dust emissions for WRF-Chem could alter the dust SWRF at TOA from cooling to warming and then lead to a positive net radiative forcing at TOA (Li and Sokolik, 2018). These different conclusions showed some degree of uncertainty in the coupled model simulations of dust aerosols' radiative forcing that needs to be further investigated in the future.
Dust aerosols can act not only as water-insoluble cloud condensation nuclei (CCN) (Kumar et al., 2009) but also as ice nuclei (IN) (Lohmann and Diehl, 2006) since they are referred to as ice-friendly (Thompson and Eidhammer, 2014). Therefore, activation and heterogeneous ice nucleation parameterizations (INPTs) with respect to dust aerosols were developed and incorporated into WRF-Chem to explore ACI effects as well as both ARI and ACI effects of dust aerosols in Asia (Jin et al., 2015, 2016b; Y. Zhang et al., 2015a; Su and Fung, 2018a, b; K. Wang et al., 2018). During dust storms, including the adsorption activation of dust particles played vital roles in the simulations of ACI-related cloud properties, with a 45 % increase in cloud droplet number concentration (CDNC) compared to a simpler aerosol activation scheme in WRF-Chem (K. Wang et al., 2018). More sophisticated INPTs implemented in WRF-Chem that take dust particles into account as IN resulted in substantial modifications of cloud and ice properties as well as surface meteorological variables and air pollutant concentrations in model simulations (Y. Zhang et al., 2015a; Su and Fung, 2018b). Y. Zhang et al. (2015a) determined that dust aerosols acting either as CCN or IN made model results rather different regarding radiation, T2, precipitation, and number concentrations of cloud water and ice. Su and Fu (2018b) described the ACI effects of dust as having fewer impacts on the radiative forcing than its ARI effects, and dust particles could promote (demote) ice (liquid) clouds in the middle to upper (lower to middle) troposphere over EA. With turning on both ARI and ACI effects of dust, fewer low-level clouds and more mid- and high-level clouds were detected that contributed to cooling at the Earth's surface and in the lower atmosphere as well as warming in the middle to upper troposphere (Su and Fung, 2018b). Mineral dust particles transported by the westerly and southwesterly winds from the Middle East (ME) affected the radiative forcing at TOA and the Earth's surface as well as in the ATM by the dust-induced ARI and ACI in the Arabian Sea and the Indian subcontinent, subsequently changing the circulation patterns, cloud properties, and characteristics related to the India summer monsoon (ISM; Jin et al., 2015, 2016a). Moreover, the effects of dust on precipitation are not only complex but also highly uncertain, as evidenced by several modeling investigations targeting a variety of areas in Asia (Jin et al., 2015, 2016a, b; Y. Zhang et al., 2015a; Su and Fung, 2018b). Less precipitation from model simulations including dust effects was found at EA, and dust particles acting mainly as CCN or IN influenced precipitation in a rather different way (Y. Zhang et al., 2015a). A positive response of ISM rainfall to dust particles from the ME was reported by Jin et al. (2015) and was less affected by dust storms from local sources and NWC (Jin et al., 2016b). Jin et al. (2016a) further elucidated the fact that the impacts of ME dust on ISM rainfall were highly sensitive to the imaginary refractive index of the dust setting in the model, so accurate simulations of the dust–rainfall interaction depended on more precise representation of radiative absorptions of dust in two-way coupled models. About a 20 % increase or decrease in rainfall due to the dust effects was detected in different areas over EA from the WRF-Chem simulations (Su and Fung, 2018b). However, it should be mentioned that a few studies targeting DRF of dust in Asia based on WRF-Chem simulations but without enabling aerosol–radiation feedbacks (Ashrafi et al., 2017; S. Chen et al., 2017b; Tang et al., 2018) were not included in this paper.
Along with the modeling research on the effects of dust aerosols on
meteorology, their impacts on air quality in Asia were explored using
two-way coupled models (Wang et al., 2013; Chen et al., 2014; Kumar et al.,
2014; M. M. Li et al., 2017a; Li and Sokolik, 2018). Many early modeling research
work involving two-way coupled models with dust only looked into the ARI or
direct radiative effects of dust particles, which are described as follows.
Taking a springtime dust storm from the Thar Desert into consideration in
WRF-Chem, the modeled aerosol optical depth (AOD) and Ångström exponent (as
indicators of aerosol optical properties and unique proxies for the surface
particulate matter pollution) demonstrated that turning on the ARI effects
of dust could reduce biases in their simulations, but were underestimated in
northern India (Kumar et al., 2014). Wang et al. (2013) pointed out that in EA,
including the longwave radiative effects of dust in the GRAPES-CUACE dust
model lowered relative errors of the modeled AOD by 15 % compared to
simulations only considering shortwave effects of dust. Comparisons
against both satellite and in situ observations showed that the WRF-Chem
model was able to capture the general spatiotemporal variations of the
optical properties and size distribution of dust particles over the main
dust sources in EA, such as the Taklimakan and Gobi deserts, but
overestimated AOD during summer and fall and also exhibited positive
(negative) biases in the fine (coarse) mode of dust particles (Chen et al.,
2014). Besides the ARI effects of dust, the heterogeneous chemistry on dust
particles' surface added in WRF-Chem accounted for 80 % of the net
reductions of O
In the maritime SEA region, peat and forest fires triggered by El Niño induced drought conditions and released a huge amount of smoke particles, which promoted dire air pollution problems in the downstream areas, and their ARI effects simulated by WRF-Chem enhanced radiative forcing at the TOA and the atmospheric stability (Ge et al., 2014). Ge et al. (2014) also pointed out that the ARI effects of these fires impaired (intensified) sea breeze during daytime (land breeze at nighttime) over this region so that their impacts on cloud cover could be positive or negative in different areas and time periods (day or night). Sea salt and volcanic ash are also important natural aerosols for regions near seashores as are active volcanoes and surrounding areas, but modeling studies of their ARI and ACI effects are relatively scarce in Asia. Based on WRF-Chem simulations, Kedia et al. (2019a) demonstrated that the feedbacks of sea salt aerosols impacted convective and non-convective precipitation rather variously in different areas of the Indian subcontinent. Jiang et al. (2019a, b) also used WRF-Chem with and without sea salt emissions to evaluate the effects of sea salt on rainfall in Guangdong Province of China, but unfortunately, no feedbacks were considered in the simulations. So far there has been no investigation targeting aerosol effects of volcanic ash from eruptions in Asia using coupled models.
Atmospheric pollutants from anthropogenic sources are the leading causes of heavy pollution events occurring in Asia due to the acceleration of urbanization, industrialization, and population growth in recent decades, particularly in China and India, and their ARI and/or ACI effects on meteorology and air quality have been quantitatively examined using two-way coupled models (Kumar et al., 2012a, b; Li and Liao, 2014; J. Wang et al., 2014; B. Zhang et al., 2015; M. Gao et al., 2016a; Yao et al., 2017; Z. Wang et al., 2018; Archer-Nicholls et al., 2019; Bharali et al., 2019). This modeling research work has been primarily focused on the ARI and/or ACI effects of anthropogenic aerosols, their specific chemical components (especially the light-absorbing aerosols, i.e., BC and brown carbon – BrC), and aerosols originating from different sources. The major findings are outlined as follows with respect to the effects of anthropogenic aerosol feedbacks on meteorology and air quality.
Concerning the meteorological responses, most papers treated anthropogenic
aerosols as a whole to explore their effects on meteorological variables
based on coupled model simulations by enabling ARI and/or ACI in WRF-Chem,
WRF-CMAQ, WRF-CMAQ, GRAPES-CUACE, and WRF-NAQPMS (Kumar et al., 2012b; J. Wang
et al., 2014; Z. Wang et al., 2014; H. Wang et al., 2015; B. Zhang et al., 2015; X. Zhang et al., 2018; Zhao et al., 2017;
Nguyen et al., 2019a, b; Bai et al., 2020). Generally, the main ARI effects
of anthropogenic aerosols resulted in decreases in SWRF, T2, WS10, and
PBLH, as well as increases in surface relative humidity (RH2) and
temperature in the ATM, which further suppressed PBL development (Y. Gao et
al., 2015; Xing et al., 2015c; M. M. Li et al., 2017b; Zhang et al., 2018; Nguyen
et al., 2019a, b). H. Wang et al. (2015) utilized GRAPES-CUACE with ARI to
study a summer haze case in the NCP area and discovered that the ARI effects
made the subtropical high less intense (
Hitherto there were several attempts to ascertain the effects of different
chemical components of anthropogenic aerosols on meteorology in Asia (Huang
et al., 2015; Ding et al., 2016, 2019; J. Gao et al., 2018; Z. Wang et al.,
2018; Archer-Nicholls et al., 2019). First of all, Asia is the region in
the world with the highest BC emissions due to burning of a large amount of
fossil fuels and biomass, and this has increasingly attracted many
researchers to probe into the ARI and/or ACI effects of BC (IPCC, 2013). As
the most important absorbing aerosol, BC induced the largest mean DRF at the TOA (positive), in the ATM (positive), and at the surface (negative) over China during 2006 (Huang et al., 2015). Ding et al. (2016) and Z. Wang et al. (2018) further applied WRF-Chem with feedbacks to
investigate how aerosol–PBL interactions involving BC suppressed the PBL
development, which deteriorated air quality in Chinese cities and was
described as the “dome effect” (namely, BC warms the atmosphere and cools the
surface, suppresses the PBL development, and eventually results in more
accumulation of pollutants). This dome effect of BC promoted
advection–radiation fog and fog–haze formation in the YRD area through
altering the land–sea circulation pattern and increasing the moisture level
(Ding et al., 2019). J. Gao et al. (2018) also pointed out that BC in the ATM
modified the vertical profiles of heating rate and equivalent potential
temperature in Nanjing, China. In India, the ARI effects of BC enhanced
convective activities, meridional flows, and rainfall in northeastern India
during the pre-monsoon season but could either enhance or suppress
precipitation during the monsoon season in different parts of the Indian
subcontinent (Soni et al., 2018). Moreover, the ARI effects of BC on surface
meteorological variables were larger than its ACI effects in EC
(Archer-Nicholls et al., 2019; Ding et al., 2019). Besides BC, the BrC
portion of organic aerosol (OA) emitted from agriculture residue burning
(ARB) was included in WRF-Chem with the parameterization scheme suggested
by Saleh et al. (2014), and the model simulations in EC revealed that at the
TOA, the net DRF of OA was
ARB is a common practice in many Asian countries after harvesting and before planting, and it can deteriorate air quality quickly as one of the most important sources of anthropogenic aerosols, so it has been attracting much attention among the public and scientists worldwide (Reid et al., 2005; Koch and Del Genio, 2010; J. Chen et al., 2017; Yan et al., 2018; Hodshire et al., 2019). Recently, the effects of ARB aerosols on meteorology has been widely explored using the two-way coupled model (WRF-Chem) in many Asian countries and regions, such as EC (Huang et al., 2016; Wu et al., 2017; Yao et al., 2017; M. Li et al., 2018), southern China (SC) (Huang et al., 2019), and South Asia (SA) (Singh et al., 2020). In general, when ARB occurred, the WRF-Chem simulations from all the studies showed that the changes in radiative forcing induced by ARB aerosols were greater than by those from other anthropogenic sources, especially in the ATM. Also, all the modeling studies indicated that ARB aerosols reduced (increased) radiative forcing at the surface (in the ATM), cooled (warmed) the surface (the atmosphere), and increased (decreased) atmospheric stability (PBLH). Furthermore, the WRF-Chem simulations with ARI demonstrated that light-absorbing carbonaceous aerosols (CAs) from ARB caused daytime (nighttime) precipitation to decrease (increase) over Nanjing in EC during a post-harvest ARB event (Huang et al., 2016). Yao et al. (2017) pointed out that their WRF-Chem simulations in EC exhibited a larger direct radiative effect (DRE) induced by BC from ARB at the TOA than previous studies. Lately, several modeling studies using WRF-Chem targeted the effects of ARI and both ARI and ACI due to ARB aerosols from countries in the Indochina, SEA, and SA regions during the planting and harvesting time (Zhou et al., 2018; Dong et al., 2019; Huang et al., 2019; Singh et al., 2020). Zhou et al. (2018) investigated how ARB aerosols from SEA mixed with mineral dust and other anthropogenic aerosols while being lifted to the middle to lower troposphere over the source region and transported to the YRD area and then affected meteorology and air quality there. The influences of ARI and ACI caused by ARB aerosols from Indochina were contrary: namely, the ARI (ACI) effects made the atmosphere over SC warmer (cooler) and drier (wetter), and the ARI effects hindered cloud formation and suppressed precipitation there (Huang et al., 2019). Dong et al. (2019) found that the warming ARI effects of ARB aerosols were smaller over the source region (i.e., SEA) than the downwind region (i.e., SC) with cloudier conditions. Annual simulations regarding the ARI effects of ARB aerosols from SA (especially Myanmar and Punjab) indicated that CAs released by ARB reduced the radiative forcing at the TOA but did not change the precipitation processes much when only the ARI effects were considered in WRF-Chem (Singh et al., 2020).
Besides ARB, to our best knowledge, there were only a few research works
quantitatively assessing the effects of anthropogenic aerosols from
different emission sources on meteorology using WRF-Chem. M. Gao et al. (2018b)
evaluated the responses of radiative forcing in China and India to aerosols
from five emission sectors (power, industry, residential, BB, and
transportation) and found that the power (residential) sector was the
dominant contributor to the negative (positive) DRF at the TOA over both
countries due to high emissions of sulfate and nitrate precursors (BC), and
the total sectoral contributions were in the order of power
In terms of anthropogenic aerosol effects on air quality, the responses of
PM
With reference to the feedback effects of anthropogenic aerosol compositions
on air quality, most modeling research work with WRF-Chem has focused on the
ARI and ACI effects of BC and BrC, especially the dome effect that
prompted the accumulation of pollutants (aerosols and O
Studies on the feedback effects of aerosols from different emission sectors
on air quality were relatively limited and mainly involved with ARB
emissions and assessments of emission controls during certain major air
pollution events. Jena et al. (2015) applied WRF-Chem with aerosol feedbacks
and investigated O
Poor air quality poses risks to human health (Brunekreef and Holgate, 2002;
Manisalidis et al., 2020); therefore, in the past several decades, air
quality models have been used in epidemiology-related research to establish
quantitative relationships between concentrations of various pollutants and
the burden of disease (including mortality and/or morbidity) as well as
associated economic loss (Conti et al., 2017). In Asia, there were several
studies that applied coupled air quality models with feedbacks to assess
human health effects of air pollutants under historical and future scenarios
(M. Gao et al., 2015, 2017c; Ghude et al., 2016; Xing et al., 2016; Wang et
al., 2017; Conibear et al., 2018a, b; Hong et al., 2019; Zhong et al.,
2019). By applying WRF-Chem with ARI and ACI, M. Gao et al. (2015) estimated
the health and financial impacts induced by an intense air pollution event
that happened in the NCP area during January 2013 and concluded that the
mortality, morbidity, and financial losses over the Beijing area were USD 690, 69 070,
and 253.8 million, respectively. Targeting the same case, Gao et al. (2017c) pointed out that turning on the data assimilation of surface
PM
Even though there are a certain number of research papers using two-way coupled models to quantify the effects of aerosol feedbacks on regional meteorology and air quality in Asia, model performances impacted by considering aerosol effects varied to some extent. This section provides a summary of model performance by presenting the SIs of meteorology and air quality variables as shown in Table S2. These SIs were collected from the selected papers supplying these indices and defined as papers with SIs (PSIs) (listed in Tables B2–B3 of Appendix B). As mentioned in Sect. 3, investigations of ACI effects were very limited, and there were no former studies simultaneously exploring aerosol feedbacks with and without both ARI and ACI turned on. Here, we only compared the SIs for simulations with and without ARI in the same study, as summarized in Appendix Tables B4–B5. It should be pointed out that all the reported evaluation results either from individual models or inter-model comparison studies were extracted and put into Table S2.
With certain emissions, accurate simulations of meteorological elements are critical to air quality modeling and prediction (Seaman, 2000; Bauer et al., 2015; Appel et al., 2017; Saylor et al., 2019). Targeting meteorological variables, we summarized their SIs and further analyzed the variations of SIs on different simulated timescales and among multiple models.
Figure 3 shows the compiled statistical indicators (correlation coefficient
–
The evaluations for T2 (Fig. 3a) from PSIs revealed that in Asia coupled
models performed rather well for temperature (mean
Figure 3b and c illustrate that RH2 was simulated reasonably well (mean
Compared with the correlation coefficients of T2, RH2, and SH2, mean
Besides the SIs discussed above, very limited papers reported the normalized
mean error (NME) (%) of surface meteorological variables (T2, SH2, RH2,
and WS10) simulated by two-way coupled models (WRF-Chem and WRF-CMAQ) in
Asia, which is summarized in Table B7. The evaluations with two-way
coupled models in Asia showed that the overall mean percent errors of T2,
SH2, RH2, and WS10 were 22.71 %, 10.32 %, 13.94 %, and 51.28 %,
respectively. The ranges of NME (%) values were quite wide for T2 (from
Quantile distributions of
Also, to examine how different coupled models (i.e., WRF-Chem, WRF-CMAQ, WRF-NAQPMS, GRAPES-CUACE, and GATOR-GCMOM) performed in Asia with respect to meteorological variables, the SIs were extracted from PSIs in terms of these five coupled models and displayed in Fig. 4. The SIs for T2, RH2, SH2, and WS10 from WRF-NAQPMS, GRAPES-CUACE, and GATOR-GCMOM simulations were missing or had rather limited samples so that the discussion here only focuses on the WRF-Chem and WRF-CMAQ simulations. Moreover, the SIs sample size from studies involving WRF-Chem was generally larger than that involving WRF-CMAQ, except for SH2.
As seen in Fig. 4a, the modeled T2 by both WRF-CMAQ and WRF-Chem was well
correlated with observations, but WRF-CMAQ (mean
Both WRF-Chem and WRF-CMAQ performed better for SH2 (mean
The modeled WS10 by both WRF-Chem and WRF-CMAQ (Fig. 4d) correlated with
observations on the same level with the mean
In general, WRF-CMAQ performed better than WRF-Chem for T2 but worse for humidity (RH2 and SH2), and both models' performance for WS10 was very similar. WRF-Chem overestimated T2, RH2, and WS10 and underestimated SH2 slightly, while WRF-CMAQ overpredicted humidity and WS10 but underpredicted T2. Compared to WRF-Chem and WRF-CMAQ, the very few SIs samples indicated that for the meteorological variables excluding SH2, WRF-NAQPMS simulations matched observations better than GRAPES-CUACE simulations, but more applications and statistical analyses of these two models are needed to make this kind of comparison conclusive.
Quantile distributions of the statistical indices for simulated surface meteorological variables by WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM in Asia.
The results of the overall statistical evaluation for the online air quality
simulations are presented in Fig. 5, and all labels and colors indicate that
the SIs are the same as those for meteorological variables. In this figure and
following figures, NP and NS are the number of publications and samples with SIs,
respectively, and are summed up in Table B3. In Fig. 5a, the correlation
between the simulated and observed PM
Compared with PM
In addition to the SIs analyzed above and similar to the surface
meteorological variables, the NME (%) of PM
Quantile distributions of statistical indices for simulated
PM
Figure 6 shows the SIs for PM
Quantile distributions of R, MB, and RMSE of PM
Aerosol feedbacks not only impact the performances of two-way coupled models
but also the simulated meteorological and air quality variables to a certain
extent. In this section, we collected and quantified the variations (Table S3) of these variables induced by ARI and/or ACI from the modeling studies
in Asia. Due to limited sample sizes in the collected papers, the target
variables only include radiative forcing, surface meteorological parameters
(T2, RH2, SH2, and WS10), PBLH, cloud, precipitation, and PM
With regard to radiative forcing, most studies with two-way coupled models
in Asia focused on the effects of dust aerosols (Dust), BC emitted from
ARB (ARB_BC) and anthropogenic sources (Anthro_BC), and total anthropogenic aerosols (Anthro). Figure 7 presents the
variations of simulated SWRF and LWRF at the bottom of the atmosphere (BOT), TOA, and in the
ATM due to aerosol feedbacks, and detailed information on these variations
is compiled in Table S5. In this figure, the color bars show the range of
radiative forcing variations, and the black tick marks inside the color bars
represent these variations extracted from all the collected papers. It
should be noted that in this figure all the radiative forcing variations
were plotted regardless of temporal resolutions of data reporting and
simulation durations. Apparently in Asia, most studies targeted the SWRF
variations induced by anthropogenic aerosols at the BOT that exhibited the
largest differences ranging from
Considering BC from anthropogenic sources and ARB, they both led to positive
SWRF at the TOA (with mean values of 2.69 and 7.55 W m
Variations of shortwave and longwave radiative forcing (SWRF and LWRF) simulated by two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) with aerosol feedbacks at the bottom and top of the atmosphere (BOT and TOA) as well as in the atmosphere (ATM) in Asia.
As shown in Fig. 7, SWRF variations at the BOT caused by total aerosols (sum
of Anthro, Anthro_BC, ARB_BC, and dust) have
been widely assessed in Asia. Therefore, we further analyzed their
spatiotemporal distributions and inter-regional differences, which are
displayed in Fig. 8. Figure 8a presents the SWRF variations over different
areas of Asia (the acronyms used in Fig. 8 are listed in Table B1)
at different timescales. In Asia, almost 41 % of the selected papers
investigated SWRF in terms of its monthly variations, 36 % its hourly
and daily variations, and 23 % its seasonal and yearly
variations. Most studies reported that aerosol-induced SWRF variations
primarily occurred in NCP, EA, China, and India. At the hourly scale, the
range of SWRF decrease was from
To identify the differences of aerosol-induced SWRF variations between high-
(Asia) and low-polluted regions (Europe and North America), their
inter-regional comparisons are depicted in Fig. 8b. This figure does not
include information about temporal resolutions of data reporting and
durations of model simulations with ARI and/or ACI, but it intends to delineate
the range of SWRF changes due to aerosol feedbacks. The SWRF variations
fluctuated from
Responses of shortwave radiation forcing to aerosol feedbacks in
different areas and periods in Asia
The impact of aerosols on radiation can influence the energy balance, which eventually alters other meteorological variables. The summary of aerosol-induced variations of T2, WS10, RH2, SH2, and PBLH in different regions of Asia as well as at different temporal scales is provided in Table 6. In this table, the minimum and maximum values were collected from the corresponding papers, and the mean values were calculated by adding all the variations from these papers and then dividing by the number of samples.
Overall, aerosol effects led to decreases in T2, WS10, and PBLH with average
changes of
Summary of variations of surface meteorological variables and planetary boundary layer height (PBLH) caused by aerosol feedbacks simulated by two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) in different regions of Asia and at different temporal scales.
In the included publications, only a few papers focused on the effects of aerosol feedbacks on cloud properties (cloud fraction, LWP, ice water path - IWP, CDNC and cloud effective radius) and precipitation characteristics (amount, spatial distribution, peak occurrence, and onset time) using two-way coupled models in Asia, as shown in Table 7. In this table, the abbreviations representing aerosol emission sources (dust, ARB_BC, Anthro_BC, and Anthro) and regions in Asia are defined in Table B1. The plus and minus signs indicate increase and decrease, respectively.
The variations of cloud properties and precipitation characteristics induced by ARI and/or ACI are rather complex and not uniform in different parts of Asia and time periods. BC from both ARB and anthropogenic sources reduced cloud fraction through ARI and both ARI and ACI in several areas in China. ARI and/or ACI induced by anthropogenic aerosols could increase or decrease cloud fraction and affect cloud fraction differently in various atmospheric layers and time periods. Considering EA and subareas in China, anthropogenic aerosols tended to increase LWP through ARI and ACI as well as ACI alone but decrease LWP in some areas of SC (ARI and ACI) at noon and in the afternoon during summertime and NC (ACI) in winter. ARI and ACI induced by anthropogenic BC aerosols had negative effects on LWP except during daytime in CC. Dust aerosols increased both LWP and IWP through ACI in EA, which was reported only by one study. The increase (decrease) in CDNC caused by the ARI and ACI effects of anthropogenic (anthropogenic BC) aerosols in EC during summertime was reported. Through ACI, anthropogenic aerosols affected CDNC positively in EA and China. Compared to anthropogenic aerosols, dust aerosols could have much larger positive impacts on CDNC via ACI in springtime over EA. The ACI effects of anthropogenic aerosols reduced the cloud effective radius over China (January) and EA (July).
Among all the variables describing cloud properties and precipitation
characteristics, the variations of precipitation amount were studied the
most using two-way coupled models in Asia. How turning on ARI and/or ACI in
coupled models can change precipitation amount is not unidirectional and
depends on many factors, including different aerosol sources, areas,
emission levels, atmospheric humidity, precipitation types, seasons, and
time of a day. Under high emission levels as well as at slightly
different humidity levels of RH
Summary of changes in cloud properties and precipitation characteristics due to aerosol feedbacks simulated by two-way coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) in Asia.
Continued.
Continued.
Note – SEC: southeastern China, EC: eastern China, CC: central China, SC: southern China, NC: northern China, SK: South Korea, PRD: Pearl River Delta, EA: East Asia, SWC: southwestern China, MRYR: Middle reaches of the Yangtze River, IGP: Indo-Gangetic Plain, WG: western Ghats, NEI: northeastern India, NI: northern India, CI: central India, NWI: northwestern India, SPI: southern peninsula of India, EI: eastern India, EPI: eastern peninsula of India, WPI: western peninsula of India, CPI: central peninsula of India, NCI: northern central India, SI: southern India.
Aerosol effects not only gave rise to changes in meteorological variables
but also air quality. Table 8 (the minimum, maximum, and mean values were
defined in the same way as in Table 6) summarizes the variations of
atmospheric pollutant concentrations induced by aerosol effects in different
regions of Asia and at different timescales. In Asia, most modeling studies
with coupled models targeted the impacts of aerosol feedbacks on surface
PM
Simulation results showed that turning on aerosol feedbacks in coupled
models generally made PM
In addition to PM
Compilation of aerosol-induced variations of PM
A schematic diagram depicting aerosol–radiation–cloud interactions and quantitative effects of aerosol feedbacks on meteorological and air quality variables simulated by two-way coupled models in Asia.
Two-way coupled models have been applied in the US and Europe extensively and then in Asia due to frequent occurrences of severe air pollution events accompanied by rapid economic growth in the region. Until now, no comprehensive study has been conducted to elucidate the recent advances in two-way coupled models' applications in Asia. This paper provides a critical overview of the current status and research focuses of related modeling studies using two-way coupled models in Asia between 2010 and 2019, and it summarizes the effects of aerosol feedbacks on meteorological and air quality variables from these studies.
By systematically searching peer-reviewed publications with several scientifically based search engines along with a variety of keyword combinations and applying certain selection criteria, 160 relevant papers were identified. Our bibliometric analysis results (as schematically illustrated in Fig. 9) showed that in Asia, research activities with two-way coupled models have increased gradually in the past decade, and five two-way coupled models (WRF-Chem, WRF-CMAQ, WRF-NAQPMS, GRAPES-CUACE, and GATOR-GCMOM) were extensively utilized to explore the ARI and/or ACI effects in Asia focusing on several high aerosol loading areas (e.g., EA, India, China, and NCP) during wintertime and/or severe pollution events, with fewer investigations looking into other areas and seasons with low pollution levels. Among the 160 papers, nearly 82 % of them focused on ARI (72 papers) and both ARI and ACI effects (60 papers), but papers only considering ACI effects were relatively limited. The ARI and/or ACI effects of natural mineral dust, BC and BrC from anthropogenic sources, and BC from ARB were mostly investigated, while a few studies quantitatively assessed the health impacts induced by aerosol effects.
Meta-analysis results revealed that enabling aerosol effects in two-way
coupled models could improve their simulation and/or forecast capabilities of
meteorology and air quality in Asia, but a wide range of differences
occurred among the previous studies, perhaps due to various model
configurations (selections of model versions and parameterization schemes)
and large uncertainties related to ACI processes and their treatments in
models. Compared to the US and Europe, the aerosol-induced decrease in the
shortwave radiative forcing was larger because of higher air pollution
levels in Asia. The overall decrease (increase) in T2, WS10, PBLH, and
O
Even though noticeable progress toward the application of two-way coupled meteorology and air quality models has been made in Asia and the world during the last decade, several limitations are still presented. Enabling aerosol feedbacks leads to higher computational cost compared to offline models, but this shortcoming can be overcome with the new developments of cluster computing technology (i.e., GPU-accelerated computing and cloud computing; GPU – Graphics Processing Unit). The latest advances in the measurements and research of cloud properties, precipitation characteristics, and physiochemical characteristics of aerosols that play pivotal roles in CCN or IN activation mechanisms can guide the improvements and enhancements in two-way coupled models, especially to abate the uncertainties in simulating ACI effects. Special attention needs to be paid to assessing the accuracies of different methodologies in terms of ARI and ACI calculations in two-way coupled models in Asia and other regions. Besides the five two-way coupled models mentioned in this paper, more models capable of simulating aerosol feedbacks (such as WRF-CHIMERE and WRF-GEOS-Chem) have become available, and projects covering more comprehensive intercomparisons of these coupled models should be conducted in Asia. Future assessments of the ARI and/or ACI effects should pay extra attention to their impacts on dry and wet deposition simulated by two-way coupled models. So far, the majority of two-way coupled model simulations and evaluations have focused on episodic air pollution events occurring in certain areas, and therefore their long-term applications and evaluations are necessary; their real-time forecasting capabilities should be explored as well.
Flowchart of literature search and identification.
Lists of abbreviations and acronyms.
Continued.
The compiled number of publications (NP) and number of samples (NS) for papers providing statistical indices (SIs) of meteorological variables.
Note that “No.*” is consistent with “No.” in Table 1, and
The compiled number of publications (NP) and number of samples (NS) for papers providing statistical indices (SIs) of air quality variables.
Note that “No.*” is consistent with “No.” in Table 1, and
The compiled number of publications (NP) and number of samples (NS) for papers simultaneously providing the statistical indices (SIs) of meteorological variables simulated by coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) with and without ARI.
Note that “No.*” is consistent with “No.” in Table 1, and
The compiled number of publications (NP) and number of samples (NS) for papers simultaneously providing the statistical indices (SIs) of air quality variables simulated by coupled models (WRF-Chem, WRF-CMAQ, GRAPES-CUACE, WRF-NAQPMS, and GATOR-GCMOM) with and without ARI.
Note that “No.*” is consistent with “No.” in Table 1, and
Description of refractive indices and radiation schemes used in the WRF-Chem and WRF-CMAQ in Asia.
Summary of normalized mean error (NME) (%) of surface meteorological and air quality variables using two-way coupled models (WRF-Chem and WRF-CMAQ).
To probe the model performance of simulated T2, RH2, SH2, and WS2 at
different temporal scales, the SIs of these meteorological variables from PSIs
were grouped according to the simulation time (yearly, seasonal, monthly, and
daily) and plotted in Fig. C1. Note that the seasonal results contained SIs
values from simulations lasting more than 1 month and less than or equal
to 3 months. Here in Fig. C1, NP and NS were the number of PSIs and samples
with SIs at different timescales, respectively, and also their total values
were the same as the ones listed in Table S2. The correlations between
simulated and observed T2 (Fig. C1a) at the seasonal (mean
Given that no SIs were available for RH2 at the seasonal scale, results at
other timescales were discussed here. Figure C1b presents simulated
RH2 at the daily scale with the best correlation coefficient (mean
Lacking SIs for SH2 at the daily scale, only those at other timescales
were compared. Even though NP and NS were very limited, the modeled SH2
(Fig. C1c) exhibited especially good correlation with observations with
the mean
As seen in Fig. C1d, the modeled WS10 at the monthly scale (mean
Figure C2 depicts the SIs of simulated PM
Regarding correlation between simulated and observed O
The statistical indices of modeled meteorological variables at different temporal scales (yearly, seasonal, monthly, and daily) from past studies in Asia.
The quantile distributions of simulated PM
The related dataset can be downloaded from
The supplement related to this article is available online at:
CG, AX, XZ, and QT carried out the data collection, related analysis, figure plotting, and paper writing. HZ, SZ, GY, and MZ were involved with the original research plan and made suggestions for the paper writing.
The contact author has declared that neither they nor their co-authors have any competing interests.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors are very grateful to many researchers who provided detailed information on the two-way coupled models and related research work. The list includes but is not limited to Xueshun Chen, Zifa Wang, Yi Gao, Meigen Zhang, and Baozhu Ge (Institute of Atmospheric Physics, Chinese Academy of Sciences); Chunhong Zhou (Chinese Academy of Meteorological Sciences); Yang Zhang (Northeastern University); Mark Zachary Jacobson (Stanford University); Tianliang Zhao (Nanjing University of Information Science & Technology); Xin Huang (Nanjing University); Chun Zhao (University of Science and Technology of China); Junhua Yang and Shichang Kang (Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences); Sachin Ghude (Ministry of Earth Sciences Government of India); and Luke Conibear (University of Leeds). We would also like to express our deepest appreciation to the editor, James Allan, and two anonymous reviewers for their constructive comments and suggestions, which helped to improve the quality and readability of this article.
This study was financially sponsored by the National Key Research and Development Program of China (grant nos. 2017YFC0212304 and 2019YFE0194500), the Talent Program of Chinese Academy of Sciences, and the National Natural Science Foundation of China (grant nos. 42171142, 41771071, and 41571063).
This paper was edited by James Allan and reviewed by two anonymous referees.