Environmental Science and Pollution Research

, Volume 25, Issue 15, pp 14844–14855 | Cite as

Impact of dynamical and microphysical schemes on black carbon prediction in a regional climate model over India

  • Rohit Srivastava
  • Sherin Hassan Bran
Research Article


Aerosol concentrations and their properties strongly depend on dynamics of atmosphere. Effects of physical and dynamical parameterizations on meteorology and black carbon (BC) mass in Weather Research and Forecasting model coupled with Chemistry (WRF-CHEM) are investigated over India. Simulations are performed in ten experiments considering two boundary layer, three cumulus parameterization, and five microphysics schemes during winter and monsoon of 2008. Morrison double-moment physical parameterization, Yonsei University boundary layer parameterization with Kain-Fritsch and Grell-Freitas cumulus parameterization schemes are found suitable to simulate meteorology and BC mass over India. BC mass is found to be underestimated in almost all experiments during winter; while, BC mass is overestimated in monsoon over Ahmedabad, Delhi, and Kanpur, which suggests inefficient wet scavenging of BC in monsoon, while lower emission rate may cause differences in winter. The results will be useful in understanding parameterizations and their impact on aerosols.


Dynamical parameterization Meteorology Aerosols Black carbon Regional climate model 


Aerosols have crucial impact on the regional and global climate system. They have myriad feedbacks on climate at a variety of scales from influencing Earth-atmospheric radiation balance (radiative feedback), and affecting cloud microphysical properties and lifetime. In case of primary aerosol characterization, the interactive emission parameterizations which depend on soil properties and/or wind intensity are utilized. For example, generally, desert dust and sea salt particles are invoked using these parameterizations in the models. In addition, the existing emission inventories for other primary aerosol types are also utilized. While, secondary aerosols are not directly emitted and are produced either from gas-phase precursors or from chemical reactions among dissolved or adsorbed gases and primary aerosols. The study of Aerosol Comparisons between Observations and Models (AEROCOM) with and without harmonized emissions suggested that although the uncertainties in emissions can be large, the inter-model diversity decreased slightly after the emission harmonization but still remained large (Textor et al. 2007). The aerosol burden is largely controlled by model-specific transport, removal, chemistry, and parameterizations of aerosol microphysics and to a lesser extent by the spatio-temporal distributions of the emissions (Textor et al. 2007). For non-harmonized emissions, the standard deviation of the total aerosol burden decreased from 18 to 16 Tg, for harmonized emissions (Textor et al. 2007). Thus, other model components, such as parameterizations of physical and dynamical processes, contribute significantly to the model uncertainties.

The mechanisms affecting the aerosol concentration are new particle formation, coagulation, and source or sink terms for a particular size resulting from condensation growth. In addition, planetary boundary layer (PBL) and transport processes can have significant impact on aerosol concentration (e.g., Chen et al. 2009). The lifetime of aerosols in the atmosphere is determined by the removal processes balance against the emission and production processes. The removal processes are crucial for species which do not interact chemically (e.g., primary aerosols) as they are represented in models by only available sinks. The removal processes of aerosols are divided in two categories: ”wet” deposition (scavenging) processes in which aerosols are removed by precipitation, and ”dry” processes in which gravitational sedimentation (or gravitational settling) removes aerosols from the atmosphere. Wet deposition is the most efficient but regionally limited aerosol sink (Pruppacher et al. 1997). The uncertainty in wet deposition is augmented by the error in precipitation computation. The dry deposition of aerosol is predominantly size dependent as different mechanisms operate at different particle sizes. The timescale of dry deposition exhibits its maximum at around 100 nm (Petroff et al. 2008). For small aerosols of nanometers diameter, timescale can be orders of magnitudes smaller due to an efficient removal mechanism by Brownian diffusion. However, for larger size aerosols (size > 100 nm), the deposition is enhanced due to interception and inertial impaction mechanisms (Petroff et al. 2008) and the respective timescale of dry deposition is smaller. The timescale of aerosol dynamical processes generally varies between thousand seconds to 105 s (Pryor and Binkowski 2004). In an unstable atmospheric boundary layer, the estimated mixing timescale is around 10 min (Stull 1988). Therefore, aerosols get sufficient time to establish well-mixed conditions in the boundary layer. Such an efficient mixing throughout the atmospheric column is not possible under near-neutral and stable conditions. The turbulent transfer and dry deposition processes give rise to aerosol dynamical terms during the transport of aerosols. The timescale of turbulent transfer characterizes the aerosol transport time within turbulent air layer, while dry deposition results in addition of the transport pathway within the laminar air layer surrounding the collecting surfaces.

The uncertainties in the meteorological fields have significant impacts on aerosols simulations in the regional climate model. Transport of aerosols is controlled by winds. Winds can also influence the interactive emission parameterizations. The cloud coverage, rain localization, and intensity can be influenced by humidity. Humidity is also crucial for hygroscopic particle growth, which is important for the particle settling and wet deposition processes. There are significant uncertainties associated to the physical and dynamical interactions with chemistry in regional climate models. The suitable physical and dynamical schemes over the Indian region are not specified in the regional climate model. In order to investigate the role of physical and dynamical parameterizations on meteorology and aerosol characteristics, Weather Research and Forecasting model coupled with Chemistry (WRF-CHEM) is utilized in the present study.


The simulations of Weather Research and Forecasting model (Skamarock et al. 2008) coupled with Chemistry (WRF-CHEM) version 3.6.1 (Grell et al. 2005; Fast et al. 2006) are investigated in the present study. The simulation domain contains 15 – 30 N and 68 – 84 E comprising western, north, and central Indian regions during winter (December-January-February) and monsoon (June-July-August-September) seasons of 2008. The simulation experiments are performed at horizontal grid resolution of 30 km × 30 km. Fifty vertical levels are considered in which 20 vertical divisions are taken within 10 km as aerosols are mainly confined in lower troposphere. The 6-h initial and lateral boundary conditions for the meteorological fields at spatial resolution of 1× 1 for year 2008 are obtained from National Centre for Environmental Predictions, Final Analysis (NCEP/FNL) and utilized in the simulations. The surface process is represented by Noah land surface (Tewari et al. 2004) and revised MM5 Monin-Obukhov scheme (Jiménez et al. 2012). The short- and long-wave radiation computation is employed by Rapid Radiative Transfer Model for General Circulation Models (RRTMG) (Iacono et al. 2008).

The chemical mechanism for gas-phase chemistry in WRF-CHEM simulations is provided from MOZART4 (Emmons et al. 2010) and for aerosol process based on Goddard Chemistry Aerosol Radiation and Transport (GOCART) bulk aerosol scheme (MOZCART) (Pfister et al 2011). MOZCART aerosols and its optical and heterogeneous properties have been tested and evaluated using WRF-Chem over Indian region (Feng et al. 2016). Externally mixed aerosol species are included in MOZCART scheme such as sulfate, black carbon, organic carbon, seven size bin (0.1–0.18, 0.18–0.3, 0.3–0.6, 0.6–1, 1–1.8, 1.8–3, and 3–6 μ m) dust and four size bin (0.1–0.5, 0.5–1.5, 1.5–5, and 5–10 μ m) sea salt.

The biogenic emissions of trace species are obtained from Model of Emissions of Gases and Aerosols from Nature (MEGAN) (Guenther et al. 2006). Emissions from open biomass burning are obtained from the Fire Inventory from NCAR version1 (FINNv1) (Wiedinmyer et al. 2011) with the addition of online plume rise module (Freitas et al. 2007). Anthropogenic emissions for black carbon (BC), organic carbon (OC), sulfate (SO4), nitrate (NO4), carbon monoxide (CO), particulate matter < 2.5 μ m (PM2.5), particulate matter < 10 μ m (PM10), ammonium (NH4), methane (CH4), and non-methane volatile organic compounds (NMVOC) are taken from Emission Database for Global Atmospheric Research collaboratively with Task Force Hemispheric Transport of Air Pollution (EDGAR-HTAP) at global horizontal grid resolution of 0.10× 0.10. The emission inventory includes the emission sources via industrial waste, land transportation, air transportation, and shipping, domestic, and agricultural fields. Photolysis and dry deposition of gases are parameterized based on Wesely (1989) and obtained from (

In the present study, ten simulation experiments are performed with different boundary layer, cumulus and microphysical parameterizations. The schemes used in different experiments are mentioned in Table 1. The brief descriptions of the utilized schemes are following.
Table 1

Simulation experiments using different boundary layer, cumulus, and microphysical schemes


Boundary layer schemes

Cumulus parameterization scheme

Microphysics scheme

Exp 1

Yonsei University (YSU)

Kain-Fritsch (KF)

Morrison double-moment

Exp 2

Yonsei University (YSU)

Grell-3D (G3)

Morrison double-moment

Exp 3

Yonsei University (YSU)

Grell-Freitas (GF)

Morrison double-moment

Exp 4

Mellor-Yamada-Janjic (MYJ)

Kain-Fritsch (KF)

Morrison double-moment

Exp 5

Mellor-Yamada-Janjic (MYJ)

Grell-Freitas (GF)

Morrison double-moment

Exp 6

Mellor-Yamada-Janjic (MYJ)

Grell-3D (G3)

Morrison double-moment

Exp 7

Yonsei University (YSU)

Grell-Freitas (GF)

New Thompson et al.

Exp 8

Yonsei University (YSU)

Grell-Freitas (GF)

WRF double-moment 5-class scheme

Exp 9

Yonsei University (YSU)

Grell-Freitas (GF)

WRF double-moment 6-class scheme

Exp 10

Yonsei University (YSU)

Grell-Freitas (GF)

Thompson aerosol-aware

Planetary boundary layer (PBL) schemes

In the forecasting model PBL schemes are used to parameterize the turbulent vertical fluxes of momentum, heat, and moisture within the PBL, and throughout the atmosphere. The turbulent fluxes are estimated from the mean quantities which require a closure scheme (Holt and Raman 1988). In local closure scheme, the turbulent fluxes are estimated at each model grid point from the mean atmospheric variables and/or their gradients at that point. The assumption, on solely dependence of fluxes on local values and gradients of basic model variables, is least valid under convective conditions (Stull 1984). In the convective condition, the turbulent fluxes transport fluid to longer distances, which is dominated by large eddies (Troen and Mahrt 1986). The non-local fluxes may be incorporated as parameterized non-local term (Troen and Mahrt 1986) or treated explicitly (Stull 1984).


This is a local closure model. The 1.5-order (level 2.5) turbulence closure model of (Mellor and Yamada 1982) was utilized in MYJ PBL scheme to represent turbulence above the surface layer (Janjiċ 1994). The MYJ scheme estimates eddy diffusion coefficients from prognostically calculated turbulent kinetic energy. (Mellor and Yamada 1982) suggested that the scheme is suitable for all stable and slightly unstable flows; however, the errors are more likely to increase as the flow approaches the free-convective condition.

Yonsei University PBL

In YSU PBL scheme, a revised vertical diffusion package with a non-local turbulent mixing coefficient was proposed. The major component of the scheme was the inclusion of an explicit treatment of entrainment processes at the top of the PBL (Hong and Noh 2006). This scheme enhances the boundary layer mixing in the thermally induced free convection regime, while decreases mixing in the mechanically induced forced convection regime, which has alleviated the common problems related to PBL in the Medium-Range Forecasting (MRF) (Hong and Noh 2006). The excessive mixing under strong wind condition in the mixed layer was resolved (Hong and Noh 2006). YSU PBL scheme has been successfully implemented in WRF model which provides a more realistic structure and development of the PBL.

Cumulus parameterizations scheme

The cumulus parameterization is a formulation of the statistical effects of moist convection to obtain a closed system for predicting weather and climate (Arakawa 2004). The cumulus parameterizations are mainly concerned with the statistical behavior of cumulus clouds under different conditions. This parameterization is similar to that of climate dynamics, which are associated with time and space means, forced, and free vibrations around means, and interactions between different temporal and spatial scales, etc (Arakawa 2004). In cumulus parameterization, the statistical effects of cumulus convection are formulated without predicting individual clouds. Due to its closure behavior, limited number of equations that govern the statistics of a system with huge dimensions are examined. Thus, cumulus parameterization, as distinguished from the dynamics and thermodynamics of individual clouds, is in the choice of appropriate closure assumptions.


The KF scheme is a mass flux parameterization which utilizes the Lagrangian parcel method, including vertical momentum dynamics (Kain 2004, and references cited therein). In this scheme, the existence of moist-convective instability is estimated, and it is examined that any existing instability will be available for cloud growth. In addition, the properties of any convective clouds are also investigated.

There are three parts in KF scheme : (i) the convective trigger function in which the scheme has to identify potential source layers for convective clouds, i.e., updraft source layers, (ii) the mass flux formulation in which convective updrafts are represented using a steady-state entraining-detraining plume model, and (iii) the closure assumption in which the mass in a column is rearranged using the updraft, downdraft, and environmental mass fluxes until at least 90% of the convective available potential energy is removed (Kain 2004). A study by Kumar et al. (2012) suggested that summer rainfall over India is dominated by the convective precipitation and the overestimation in rainfall by the model could be due to use of the KF scheme.


The cumulus parameterizations differ fundamentally in closure assumptions and parameters used to solve the interaction (between the larger scale flow and the convective clouds) problem, which lead to a large spread and uncertainty in possiblesolutions (Grell and Dévényi 2002). G3 scheme utilizes these uncertainties by combining ensemble and data assimilation techniques. A parameterization, which can employ a large ensemble of closure assumptions and parameters, is developed (Grell and Dévényi 2002). The closures and parameters are obtained from the cumulus parameterizations which are currently used in different three-dimensional models. Proper feedback to the three-dimensional model is applied utilizing the statistical techniques and ensemble probability density function (Grell and Dévényi 2002). Ojha et al. (2016) found the simulated water vapor vertical profiles using G3 scheme showed good agreement with the measured profiles which suggest G3 scheme can be reasonable convection over the Indian subcontinent.


In traditional convective parameterization scheme, the convections have been designed to be self-contained within one grid column. The fraction of the grid column, that is occupied by active convection, is assumed to be small (Grell and Freitas 2014). However, such assumption starts to break down with reducing the horizontal grid spacing (Grell and Freitas 2014). GF cumulus convective parameterization addresses the almost cloud-resolving scales (gray scales) in which horizontal resolution is < 10 km. In addition, GF also considers transport of chemical constituents and possible interactions with aerosols (Grell and Freitas 2014).

Microphysics scheme

Microphysical processes control formation of cloud droplets and ice crystals, their growth and fallout as precipitation. Microphysics plays a crucial role in regional climate models. The main microphysical processes are latent heat cooling, condensate loading, precipitation, coupling with surface processes, radiative transfer, and cloud-aerosol-precipitation interactions. There are two different approaches in describing cloud microphysics in atmospheric models, (i) the explicit bin-resolving method and (ii) the bulk method.

In bin-resolving microphysics models, particle size distribution is explicitly estimated and therefore provides more rigorous solutions. However, the bin microphysics is computationally expensive, therefore remains a restriction for climate prediction. Bulk microphysical parameterization scheme is relatively simple which estimates several drop size distribution moments rather than the drop size distribution itself (Lim and Hong 2010).

Morrison double-moment scheme

This scheme is utilized to investigate the formation and evolution of trailing stratiform precipitation under an idealized two-dimensional squall line condition (Morrison et al. 2009). The two-moment microphysics scheme was implemented in a regional climate model to predict the mass mixing ratios and number concentrations of five hydrometeor species, viz, cloud droplets, cloud ice, snow, rain, and graupel (Morrison et al. 2009).

The new Thompson et al. scheme

In this scheme, bulk microphysical parameterization (BMP) has been improved to both physical processes and computer coding. It employs many techniques found in far more sophisticated spectral/bin schemes using lookup tables (Thompson et al. 2008). It was assumed that snow size distribution depends on both ice water content and temperature, and was represented as a sum of exponential and gamma distributions (Thompson et al. 2008). In addition, based on different observations, the snow was considered as a non-spherical shape with a bulk density that varies inversely with diameter; however, other BMPs assume spherical snow with constant density (Thompson et al. 2008).

WRF double-moment 6-class scheme

A double-moment bulk cloud microphysics scheme is based on single-moment 6-class (WSM6) microphysics scheme. The mixing ratios of six water species (water vapor, cloud droplets, cloud ice, snow, rain, and graupel) in the WSM6 scheme have been estimated. In addition, the number concentrations for cloud and rainwater are also predicted in the WDM6 scheme, together with a prognostic variable of cloud condensation nuclei (CCN) number concentration (Lim and Hong 2010).

WRF double-moment 5-class scheme

The double-moment 5-class (WDM5) scheme has been developed by adding the number concentrations for warm-phase physics to the single-moment 5-class (WSM5) scheme. This simplify the computations and theoretically based on the fact that the prediction of ice-phase number concentrations has significantly less impact on the results, when compared to the prediction of warm-phase concentrations in deep convective cases (Morrison et al. 2009).

Thompson aerosol-aware scheme

Aerosols play crucial role in cloud microphysics by nucleating cloud and ice particles. An increase in aerosol concentration generally leads to more numerous, but smaller droplets for a given liquid water content (Twomey 1974). In this scheme, bulk microphysical parameterizations with explicit cloud droplet nucleation and ice activation by aerosols are implemented (Thompson and Eidhammer 2014). The schemenucleates water and ice from their dominant respective nuclei and tracks and predicts the number of available aerosols.

Results and discussion


Taylor diagram comparing rainfall in different experimental simulations of WRF-CHEM against IMD rainfall observations during monsoon (June-September) season of 2008 is shown in Fig. 1. In the Taylor diagram, the statistics are presented by correlation coefficient, normalized standard deviation (SD), and normalized root mean square difference (NRMSD). The SD and RMSD are normalized using the SD of the observational data. In the diagram, ”OBS” point represents the observational data which is used to compare the simulations in different experiments.
Fig. 1

Taylor diagram comparing the observed rainfall with different simulation experiments. The observed rainfall is obtained from IMD 0.25 gridded rainfall data. Different model experiments are mentioned in Table 1

Daily gridded rainfall data of India Meteorological Department (IMD) at 0.25× 0.25 grid resolution (Pai et al. 2014) are utilized in the study for comparison. The 0.25 gridded rainfall data was developed using daily rainfall records from the quality-controlled data of 6955 rain gauge stations over India (Pai et al. 2014). The gridded data is developed using the simplest form of inverse distance -weighted interpolation (IDW) scheme (Pai et al. 2014).

The rainfall in experiments 1, 3, and 4 shows correlation > 0.75 with the observations. The normalized root mean square (NRMS) differences between rainfall in experiments 3, 4, 5, 9, 10, and observations are less than 1. However, in experiments 1 and 7, the NRMS deviations are > 1.0. In experiments 2, 6, and 8, the NRMS differences are even greater than 1.5. The rainfall in experiments 3, 4, and 5 is found to agree with that of the observations. The higher NRMS difference results the more dry or wet biases in the simulations. The aerosol near surface are removed by wet scavenging during monsoon season. The dry bias in the simulations may lead to reduction in wet scavenging of aerosols. This suggests that the Morrison double-moment microphysical scheme with MYJ and YSU boundary layer schemes has performed better rainfall variability when compared to other schemes. The experiment 3 shows very high values of rainfall in a western Indian (part of Rajasthan and Gujarat) region, while in observations, this region receives lowest rainfall (Fig. 2).
Fig. 2

Spatial variation of rainfall in different simulation experiments during monsoon season of 2008. The observed rainfall is from IMD 0.25 gridded rainfall data

The spatial variations of rainfall in experiments 1, 2, and 6 show wet bias (double rainfall values) over central, eastern, and northern regions (Fig. 2). The rainfall over western India is overestimated in experiments 2, 3, 7, and 9. The Yonsei University (YSU) boundary layer scheme with Grell-Freitas (GF) cumulus scheme has shown overestimation of rainfall in western India for all the microphysical schemes. High rainfall in all the experiments is simulated over the western ghat region, which is consistent with the observations. The rainfall simulations in experiments 3 and 8 show good agreement with the observations in terms of spatial pattern and absolute values as well. However, rainfall in the part of the western Indian region is higher when compared to that of observations. Thus, the combination of Mellor-Yamada-Janjic (MYJ) boundary layer scheme, Kain-Fritsch (KF) or Grell-Freitas (GF) cumulus schemes, and Morrison double-moment microphysical scheme is able to simulate rainfall over the Indian region.

Relative humidity

RH is an important factor which leads to condensation growth of hygroscopic aerosols. The Taylor diagrams for the comparison of observed RH in different simulation experiments during winter and monsoon are shown in Fig. 3. RH in all the simulations is found to be within 0.55 to 0.8 correlation except experiments 2, 4, and 8 in winter, while during monsoon, the simulated RH in experiments 1, 2, 6, 8, and 9 shows correlation < 0.55. The normalized root mean square (NRMS) differences are > 1 in winter and < 1 in monsoon for most of the experiments (Fig. 3). The absolute values of RH are closer to those of observation in monsoon when compared to winter. The normalized standard deviation is also with 0.4 to 1.0 during monsoon suggesting the RH variability in simulations is similar to that in observations. The absolute differences in RH are higher in winter suggesting the experiments are more prone to higher RH.
Fig. 3

Taylor diagram comparing observed relative humidity with RH simulated in different model experiments (Table 1) during winter (December-February) and monsoon (June-September) of 2008

Wind speed

Taylor diagram for comparing near surface wind speed between WRF-CHEM simulation experiments meteorological observations during winter (December-January-February) and monsoon seasons of 2008 is shown in Fig. 4. The surface meteorological data measured over Mumbai, Ahmedabad, Delhi, and Lucknow are obtained from IMD measurements. Wind speed simulated in experiments 1, 3, 4, 5, and 8 shows correlations > 0.55 during monsoon, while experiments 1, 3, and 10 show > 0.55 correlations in winter. NRMS deviation is < 1 for experiment 3 during winter, while NRMS differences for experiments 1, 3, 4, and 8 are between 1 and 1.5.
Fig. 4

Taylor diagram comparing observed wind speed with speed in different simulation experiments (Table 1) during winter (December-February) and monsoon (June-September) of 2008

Planetary boundary layer

PBL plays crucial role while computing BC mass concentration. The Taylor diagrams for the comparison of MERRA (Modern-Era Retrospective analysis for Research and Applications) reanalysis data in different simulation experiments during winter and monsoon seasons are shown in Fig. 5. PBL heights in experiment 3 are found to have correlation ∼ 0.5 except all other experiments in both the seasons. The normalized root mean square (NRMS) differences are < 1 in winter and > 1.5 in monsoon for most of the experiments (Fig. 5). The absolute values of PBL is closer to observation in winter when compared to those in monsoon. The normalized standard deviation is also with 0.1 to 1.0 during winter suggesting the PBL variability in simulations is similar to that in reanalysis data. Yonsei University (YSU) boundary layer scheme with cumulus parameterizations scheme of Grell-Freitas (GF) with the microphysics scheme of Morrison double-moment can be able to simulate PBL height over the region. In winter, boundary layer height is lower which provides less space to aerosols, while in monsoon, the deeper boundary layer height gives rise to more space. This is one of the reasons of lower BC mass concentration in monsoon.
Fig. 5

Taylor diagram comparing reanalysis planetary boundary layer (PBL) height with PBL height in different simulation experiments (Table 1) during winter (December-February) and monsoon (June-September) of 2008

Vertical profile of temperature, RH, and wind speed

The vertical profiles of temperature, relative humidity, and wind speed observed by radiosonde data and simulated by WRF-CHEM in different experiments during winter and monsoon are also included in supplementary material (Figs. S1 and S2 respectively). The vertical profiles of temperature, relative humidity, and wind speed at four different locations (Mumbai, Ahmedabad, Delhi, and Nagpur) are obtained from University of Wyoming. The model-simulated profiles are found to match over the locations; however, there are some differences in absolute values. Relative humidity profiles are well simulated by experiments 3 and 4 over most of the locations (Fig. S2). Wind speed above 4 km shows larger differences during winter (Fig. S2) and closed to radiosonde observations during monsoon (Fig. S1).

Suitable parameterization scheme

In all the simulation experiments (Table 1), experiment 3 is found to be closer with the meteorological observations (rainfall, winds, and RH). Therefore, boundary layer scheme of Yonsei University (YSU) and cumulus parameterizations scheme of Grell-Freitas (GF) with the microphysics scheme of Morrison double-moment is found to be suitable for western, central, and northern Indian regions. The spatial variations of RH and wind for this scheme for winter and monsoon seasons are shown in Fig. 6. The lower RH in winter, while higher in monsoon is well represented in the simulation of experiment 3. The winds are calm and northeasterly during winter, while strong and southwesterly in monsoon (Fig. 6). Southwesterly winds from the oceanic region transport moist air to the land region and result in higher RH during monsoon seasons. Thus, experiment 3 is the most suitable parameterization over the region.
Fig. 6

Spatial variation of relative humidity (%) and winds (ms− 1) in simulation experiment 3 during winter (December-February) and monsoon (June-September), 2008

Black carbon mass concentration

The comparisons of black carbon mass concentrations with observations in all the experiments over Pune, Ahmedabad, Delhi, and Kanpur during two contrasting seasons winter and monsoon of 2008 are shown in Fig. 7. The BC mass concentration data were obtained from Safai et al. (2007) over Pune, from Ramachandran and Kedia (2010) over Ahmedabad, from Singh et al. (2011) over Delhi, and Ram et al. (2010) over Kanpur. All the observed BC mass data are distributed among years 2005 – 2008. The inter-annual variability in aerosols is much less when compared to the seasonal variability (Srivastava 2017); therefore, the observed data from these years may not have any significant change in the results. BC mass in the simulations of experiments 3 and 7 – 10 shows match within ± 1 σ with the observations during winter. The simulations in experiments 1 – 2 and 4 – 6 always underestimate the BC mass when compared to those in observations. During monsoon, BC mass is overestimated in all the experiments over Ahmedabad and Kanpur. The simulations of experiments 3 and 7 – 10 show agreement between ± 1 σ; however, the simulated mass is higher than that of the observations over Ahmedabad, Delhi, and Kanpur during monsoon. The simulated mass is lower in Pune. The experiment 3 is found to be able to well explain the observed changes in BC mass concentration along with the meteorological observations over the different locations.
Fig. 7

Comparison of black carbon mass concentration between observations and all simulation experiments during winter and monsoon seasons. Please see the text for details. Observed BC mass are taken over Ahmedabad from Ramachandran and Kedia (2010), over Delhi from Singh et al. (2011), over Kanpur from (Ram et al. 2010) and Pune from Safai et al. (2007)

The spatial variations in BC mass over the western, cen- tral, and northern Indian regions during winter and monsoon seasons for the most suitable experiments (experiment 3) are shown in Fig. 8. Higher BC mass can be seen over the Indo-Gangetic basin (IGB) region when compared to that in other part of subcontinent. High mass concentration over urban centers can also be captured by the model simulations. The western Indian region exhibits lower BC concentration than the other regions. The study with another regional climate model (RegCM) also suggested the higher BC mass over the IGB and lower over the western Indian region (Srivastava and Sherin 2017). BC mass is found to be lower during monsoon when compared to that in winter season owing to the wet scavenging of aerosols due to monsoonal rain and boundary layer mixing. In winter, the lower boundary layer traps aerosol near the surface and gives rise to higher BC mass. Higher BC mass concentration was found over Ahmedabad, Delhi, and Kanpur in most of the experiments during monsoon suggesting the wet scavenging of aerosols may have uncertainties. The selection of appropriate parameterization is crucial to investigate the variability in different aerosols including BC and estimating their climatic impacts. The chemical parameterization is also important and needs to be included in further studies.
Fig. 8

Spatial variation of black carbon mass concentration in simulation experiment 3 during winter (December-February) and monsoon (June-September), 2008

Summary and conclusions

The role of physical and dynamical parameterizations on meteorology and aerosol characteristics in Weather Research and Forecasting model coupled with Chemistry (WRF-CHEM) is investigated. The ten model simulation experiments are designed considering two boundary layer schemes (Yonsei University (YSU) and Mellor-Yamada-Janjic (MYJ)); three cumulus parameterization schemes (Kain-Fritsch (KF), Grell-3D (G3), and Grell-Freitas (GF)); and five microphysics schemes (Morrison double-moment, New Thompson et al., WRF double-moment 5-class scheme, WRF double-moment 6-class scheme, and Thompson aerosol-aware). The study on parameterization over western, central, and northern Indian regions reveals the following:
  • The rainfall in monsoon season can be simulated over the region using Morrison double-moment physical parameterization scheme and YSU boundary layer scheme with KF and GF cumulus parameterization schemes.

  • Grell 3D cumulus parameterization scheme is not able to capture the variabilities in meteorological parameters (rainfall, wind, and RH) over the study region. The correlations between observations and simulations in experiments having G3 cumulus parameterization show lower values (< 0.4) in different meteorological parameters. This suggests that G3 scheme is not suitable over the study region.

  • The experiment 3 is found to be most suitable to explain the variations in meteorological parameters and BC mass over the region. This experiment considers YSU boundary layer scheme, GF cumulus parameterization, and Morrison double-moment microphysical schemes.

  • BC mass is found to underestimate in almost all the experiments during winter over Pune, Delhi, and Kanpur, while BC mass is overestimated in monsoon over Ahmedabad, Delhi, and Kanpur. This suggests that the wet scavenging is not efficient to reduce the BC mass in model simulations during monsoon season, while lower emission rate may cause differences in winter.

This study will be useful in understanding the role of different parameterizations on BC mass in a regional climate model and helpful in improving the future climate predictions.



Meteorological data for initial and boundary conditions are downloaded from The initial and boundary conditions for chemical field, biogenic emissions, biomass burnings, anthropogenic emissions and programs used to process the data sets are obtained from The Radiosonde data are obtained from University of Wyoming. Authors are grateful to Rajesh Kumar and his team (UCAR, USA) for their support in WRF-Chem installation and simulations. Authors are thankful to anonymous reviewers for their valuable comments and suggestions.

Funding information

The funding for the study is provided by Department of Science and Technology (DST), Government of India (SR/S4/AS-107/2012).

Supplementary material

11356_2018_1607_MOESM1_ESM.docx (696 kb)
(DOCX 695 KB)


  1. Arakawa A (2004) The cumulus parameterization problem: past, present, and future. J Clim 17(13):2093–2525CrossRefGoogle Scholar
  2. Chen Y, Zhao C, Zhang Q, Deng ZZ, Huang MY, Ma XC (2009) Aircraft study of mountain chimney effect of Beijing. China J Geophys Res 114(D8),
  3. Emmons LK, Walters S, Hess PG, Lamarque JF, Pfister GG, Fillmore D, Granier C, Guenther A, Kinnison D, Laepple T, Orlando J, Tie X, Tyndall G, Wiedinmyer C, Baughcum SL, Kloster S (2010) Description and evaluation of the model for ozone and related chemical tracers, version 4 (MOZART-4). Geosci Model Dev 3(1):43–67CrossRefGoogle Scholar
  4. Fast JD, Gustafson WI, Easter RC, Zaveri RA, Barnard JC, Chapman EG, Grell GA, Peckham SE (2006) Evolution of ozone, particulates, and aerosol direct radiative forcing in the vicinity of houston using a fully coupled meteorology-chemistry-aerosol model. J Geophys Res 111(D21),
  5. Feng Y, Kotamarthi V, Coulter R, Zhao C, Cadeddu M (2016) Radiative and thermodynamic responses to aerosol extinction profiles during the pre-monsoon month over South Asia. Atmos Chem Phys 16:247–264. CrossRefGoogle Scholar
  6. Freitas SR, Longo KM, Chatfield R, Latham D, Silva Dias MAF, Andreae MO, Prins E, Santos J C, Gielow R, Carvalho Jr JA (2007) Including the sub-grid scale plume rise of vegetation fires in low resolution atmospheric transport models. Atmos Chem Phys 7:3385–3398CrossRefGoogle Scholar
  7. Grell GA, Dévényi D (2002) A generalized approach to parameterizing convection combining ensemble and data assimilation techniques. Geophys Res Letts 29(14, 1693)
  8. Grell GA, Peckham SE, Schmitz R, McKeen SA, Frost G, Skamarock WC, Eder B (2005) Fully coupled “online” chemistry within the wrf model. Atmos Env 39(37):6957–6975CrossRefGoogle Scholar
  9. Grell GA, Freitas SR (2014) A scale and aerosol aware stochastic convective parameterization for weather and air quality modeling. Atmos Chem Phys 14:5233–5250CrossRefGoogle Scholar
  10. Guenther A, Karl T, Harley P, Wiedinmyer C, Palmer PI, Geron C (2006) Estimates of global terrestrial isoprene emissions using MEGAN (model of emissions of gases and aerosols from nature). Atmos Chem Phys 6:3181–3210CrossRefGoogle Scholar
  11. Holt T, Raman S (1988) A review and comparative evaluation of multilevel boundary layer parameterizations for first-order and turbulent kinetic energy closure schemes. Rev Geophys 26:761–780CrossRefGoogle Scholar
  12. Hong SY, Noh Y (2006) A new vertical diffusion package with an explicit treatment of entrainment processes. Mon Weather Rev 134:2318–2341CrossRefGoogle Scholar
  13. Iacono MJ, Delamere JS, Mlawer EJ, Shephard MW, Clough SA, Collins WD (2008) Radiative forcing by long-lived greenhouse gases: calculations with the AER radiative transfer models. J Geophys Res 113:D13. CrossRefGoogle Scholar
  14. Janjiċ ZI (1994) The step-mountain eta coordinate model: further developments of the convection, viscous sublayer, and turbulence closure schemes. Mon Weather Rev 122(5):927–945CrossRefGoogle Scholar
  15. Jiménez PA, Dudhia J, González-Rouco JF, Navarro J, Montávez JP, García-Bustamante E (2012) A revised scheme for the WRF surface layer formulation. Mon Weather Rev 140:898–918CrossRefGoogle Scholar
  16. Kain KF (2004) The Kain–Fritsch convective parameterization: an update. J Appl Meteorol 43:170–181CrossRefGoogle Scholar
  17. Kumar R, Naja M, Pfister GG, Barth MC, Brasseur GP (2012) Simulations over South Asia using the Weather Research and Forecasting model with Chemistry (WRF-Chem): set-up and meteorological evaluation. Geosci Model Dev 5:321–343CrossRefGoogle Scholar
  18. Lim KSS, Hong SY (2010) Development of an effective double-moment cloud microphysics scheme with prognostic cloud condensation nuclei (CCN) for weather and climate models. Mon Weather Rev 138:1587–1612CrossRefGoogle Scholar
  19. Mellor GL, Yamada T (1982) Development of a turbulence closure model for geophysical fluid problems. Rev Geophys 20:851–875CrossRefGoogle Scholar
  20. Morrison H, Thompson G, Tatarskii V (2009) Impact of cloud microphysics on the development of trailing stratiform precipitation in a simulated squall line: comparison of one- and two-moment schemes. Mon Weather Rev 137:991–1007CrossRefGoogle Scholar
  21. Ojha N, Pozzer A, Rauthe-Schöch A, Baker AK, Yoon J, Brenninkmeijer CAM, Lelieveld J (2016) Ozone and carbon monoxide over India during the summer monsoon: regional emissions and transport. Atmos Chem Phys 16:3013–3032CrossRefGoogle Scholar
  22. Pai DS, Latha S, Rajeevan M, Sreejith O P, Satbhai NS, Mukhopadhyay B (2014) Development of a new high spatial resolution (0.25 × 0.25) long period (1901 - 2010) daily gridded rainfall data set over India and its comparison with existing data sets over the region. Mausam 65:1–18Google Scholar
  23. Petroff A, Mailliat A, Amielh M, Anselmet F (2008) Aerosol dry deposition on vegetative canopies. Part II: A new modelling approach and applications. Atmos Environ 42:3654–3683CrossRefGoogle Scholar
  24. Pfister GG et al (2011) Characterizing summertime chemical boundary conditions for airmasses entering the US West Coast. Atmos Chem Phys 11:1769–1790. CrossRefGoogle Scholar
  25. Pruppacher HR, Klett JD, Wang PK (eds) (1997) Microphysics of clouds and precipitation. Kluwer Academic Publishers, DordrechtGoogle Scholar
  26. Pryor SC, Binkowski FS (2004) An analysis of the time scales associated with aerosol processes during dry deposition. Aerosol Sci Tech 38:1091–1098. CrossRefGoogle Scholar
  27. Ram K, Sarin MM, Tripathi S (2010) Inter-comparison of thermal and optical methods for determination of atmospheric black carbon and attenuation coefficient from an urban location in northern India. Atmos Res 97:335–342CrossRefGoogle Scholar
  28. Ramachandran S, Kedia S (2010) Black carbon aerosols over an urban region: radiative forcing and climate impact. J Geophys Res 115,
  29. Safai P, Kewat S, Praveen P, Rao P, Momin G, Ali K, Devara P (2007) Seasonal variation of black carbon aerosols over a tropical urban city of Pune, India. Atmos Env 41:2699–2709CrossRefGoogle Scholar
  30. Singh S, Gupta NC, Soni K, Tanwar RS, Nath S, Arya BC, Gere BS (2011) Variation in aerosol black carbon concentration and its emission estimates at the mega-city Delhi. Int J Remote Sens 32:6749–6764CrossRefGoogle Scholar
  31. Skamarock WC, Klemp JB, Dudhia J, Gill DO, Barker DM, Wang W, Powers JG (2008) A description of the advanced research wrf version 2. Technical report, DTIC DocumentGoogle Scholar
  32. Srivastava R (2017) Trends in aerosol optical properties over South Asia. Int J Climatol 37:371–380. CrossRefGoogle Scholar
  33. Srivastava R, Sherin SH (2017) Spatio-temporal variations of black carbon and optical properties in a regional climate model. Int J Climatol 37:1432–1443. CrossRefGoogle Scholar
  34. Stull RB (1984) Transilient turbulence theory Part I: The concept of eddy-mixing across finite distances. J Atmos Sci 41:3351–3367CrossRefGoogle Scholar
  35. Stull RB (ed) (1988) An introduction to boundary layer meteorology. Kluwer Academic Publishers, DordrechtGoogle Scholar
  36. Tewari M, Chen F, Wang W, Dudhia J, LeMone M, Mitchell K, Ek M, Gayno G, Wegiel J, Cuenca R (2004) Implementation and verification of the unified noah land surface model in the WRF modelGoogle Scholar
  37. Textor C et al (2007) The effect of harmonized emissions on aerosol properties in global models - an AeroCom experiment. Atmos Chem Phys 7:4489–4501. CrossRefGoogle Scholar
  38. Thompson G, Field PR, Rasmussen RM, Hall WD (2008) Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part II: Implementation of a new snow parameterization. Mon Weather Rev 136:5059–5115CrossRefGoogle Scholar
  39. Thompson G, Eidhammer T (2014) A study of aerosol impacts on clouds and precipitation development in a large winter cyclone. J Atmos Sci 71:3636–3658CrossRefGoogle Scholar
  40. Troen I, Mahrt L (1986) A simple model of the atmospheric boundary layer: sensitivity to surface evaporation. Bound-Layer Meteorol 37:129–148CrossRefGoogle Scholar
  41. Twomey S (1974) Pollution and the planetary albedo. Atmos Environ 8:1251–1256CrossRefGoogle Scholar
  42. Wesely ML (1989) Parameterization of surface resistances to gaseous dry deposition in regional-scale numerical models. Atmos Env 23(6):1293–1304CrossRefGoogle Scholar
  43. Wiedinmyer C, Akagi SK, Yokelson RJ, Emmons LK, Al-Saadi JA, Orlando JJ, Soja AJ (2011) The fire inventory from NCAR (FINN): a high resolution global model to estimate the emissions from open burning. Geosci Model Dev 4:625–641CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Indian Centre for Climate and Societal Impacts Research (ICCSIR)Mandvi, KachchhIndia
  2. 2.Atmospheric Research UnitNational Astronomical Research Institute of ThailandChiang MaiThailand

Personalised recommendations