Black carbon aerosols over an urban region: Radiative forcing and climate impact



[1] Black carbon (BC) mass concentration in Ahmedabad, an urban location, varies from 2 μg m−3 during summer to 11 μg m−3 during winter and postmonsoon seasons. Aerosol optical depth (AOD) is higher (0.63) in summer when compared to winter (0.31). BC mass concentrations in Ahmedabad are governed by local sources and meteorology (boundary layer, winds, rainfall, and long-range transport). Single-scattering albedo (SSA) deduced using measured BC mass concentration as input in an aerosol optical properties model varies from ∼0.7 during winter and postmonsoon to 0.93 in monsoon over Ahmedabad. Surface (SFC) and atmosphere (ATM) aerosol radiative forcing (ARF) in premonsoon and monsoon are ∼50% lower than those obtained during winter and postmonsoon despite higher AODs. ATM forcing is more positive for lower SSA and AOD, while it is less positive for higher AOD and SSA. It is shown that when the amounts of BC and water vapor are high over continental regions, the net (shortwave + longwave) ATM warming will be higher. BC aerosols alone contribute on average 60% and 25% of shortwave and longwave ATM forcing. Seasonal mean heating rates are higher than 1.5 K/d in winter and postmonsoon. Heating rates including BC aerosols are at least a factor of 3 higher than when BC aerosols are absent, thus highlighting the crucial role BC aerosols play in modifying the radiation budget and climate. Thus, it is possible that BC aerosols because of their radiative and climate impacts could be contributing to the decreasing trend in rainfall over India.

1. Introduction

[2] Black carbon (BC), among the carbonaceous aerosols, is gaining considerable significance because of its ability to influence air quality and climate on local, regional, and global scales. Black carbon aerosols are produced as primary particles from incomplete combustion processes such as fossil fuel and biomass burning, and hence most of the BC in the atmosphere originates from human-made activities. BC emissions have varied in response to changes of fossil fuel usage and technology development, and the estimated fossil fuel BC emissions are the highest in the developing countries, especially China and India [Novakov et al., 2003]. The radiative and climate impacts of BC are increasingly recognized as BC can absorb sunlight, heat the air, and contribute to global warming, unlike the other aerosol types (e.g., sulfate) which produce cooling. BC is the second strongest contributor to global warming next to carbon dioxide [Ramanathan and Carmichael, 2008]. Both absorbing (BC) and scattering (sulfate) aerosols cool the surface by reducing the incoming solar radiation and cause solar dimming. The solar dimming is a global phenomenon even though the aerosol sources are local, due to diverse geographic locations of aerosol sources, abundances of aerosols, their continuous emissions, and long-range transport. The sulfate aerosols are hygroscopic and are perfect scatterers over most of the shortwave region, while, in the longwave, sulfate aerosols act as absorbers. In contrast, BC is hydrophobic and is the largest absorber of the radiation in the atmosphere both in shortwave and longwave regions.

[3] The present study focuses on the monthly mean variations in aerosol physical and optical properties over an urban region and the resultant radiative effects with an emphasis on the black carbon aerosol due to its crucial role in regional and global climate impacts. BC aerosol mass concentrations measured over an urban location in western India during 2008 along with satellite (MODIS) retrieved aerosol optical depths are utilized in the current study. The radiative effects of urban aerosols, comprising insoluble aerosols (mainly soil particles with some amount of organics), water-soluble aerosols (sulfate, nitrates, and organics) and BC aerosols, in the shortwave and longwave regions are estimated using a discrete ordinate radiative transfer algorithm. The aerosol radiative forcings are estimated for a range of relative humidity values observed over an urban region to examine the sensitivity and determine the contribution of shortwave and longwave aerosol radiative forcings by scattering and absorbing aerosols. The estimated aerosol radiative forcings for varying BC mass concentrations are compared and contrasted with forcings estimated for no-BC conditions in the atmosphere. The climate implications in terms of the heating rates (Kelvin per day) for different scenarios of BC mass and no-BC conditions are presented.

2. Study Location, Ahmedabad: Site Description and Meteorology

[4] The study location, Ahmedabad (23.03°N, 72.5°E, 55 AMSL), is an urban, densely populated (population of about 5.8 million) and industrialized city in western India. Ahmedabad has several small- and large-scale industries and a large number of automobiles that include buses, cars, two-wheelers (motorbikes and scooters), and three-wheelers (auto rickshaws), all of which contribute significantly to the production of aerosols, including black carbon [Ramachandran and Rajesh, 2007]. The mean synoptic surface winds during winter monsoon (December–January–February) over the western region of India, where Ahmedabad is located, are calm and north-northeasterly and from the polluted Northern Hemisphere. During the southwest summer monsoon (June–July–August–September), the winds are stronger and moist and come from the marine and western regions surrounding India. The wind patterns start shifting in direction during postmonsoon (October–November) from southwest to northeast. During the premonsoon season (March–April–May), the winds originate and travel from the west of Indian subcontinent. Monthly mean temperature, relative humidity, and wind speed for Ahmedabad from January to December 2008 are plotted in Figure 1. The monthly mean temperature over Ahmedabad in winter is around 20°C, which increases to above 30°C during May. Relative humidity shows a strong seasonal variation over Ahmedabad. The monthly mean wind speeds are low (<2 m s−1) during December–March, which increases to >3 m s−1 during May–June (Figure 1c). The rainfall is restricted to the southwest monsoon over Ahmedabad.

Figure 1.

Monthly mean (a) temperature, (b) relative humidity (%), (c) wind speed (m s−1), and (d) rainfall (mm) over Ahmedabad during 2008. Vertical bars indicate ±1σ deviation from the mean.

3. Measurements and Methodology

3.1. Black Carbon Mass Concentration Measurements

[5] Black carbon mass concentration measurements using a seven-wavelength (370, 470, 520, 590, 660, 880, and 950 nm) aethalometer (Magee Scientific, USA) were made in Physical Research Laboratory (PRL), Ahmedabad. PRL is located in the western part of Ahmedabad in Navrangpura, surrounded by a number of residential and business complexes. This results in a heavy movement of vehicular traffic which has a distinct diurnal variation. Within a radius of 25 km with respect to the measurement site, small-scale industrial complexes and two coal-based thermal power plants are located. BC mass measurements are made from an altitude of about 20 m above the ground using the inlet tube and pump of the aethalometer. The aethalometer was operated at the flow rate of 3 l min−1 for 24 h/d at a time resolution of 5 min.

[6] The aethalometer measures BC mass concentrations from the attenuation of a beam of light transmitted through the sample collected on a filter, which is proportional to the amount of BC mass loading in the filter deposit [Hansen and Novakov, 1990]. The light transmission is detected using a set of two photo diodes as follows: one shines through the sample spot and the other shines through a blank portion of the filter that is called the reference spot. The yielded attenuation absorption coefficient is then converted into BC mass concentration. The conversion of attenuation absorption coefficient into BC mass concentration is done using appropriate absorption efficiency, which varies as a function of wavelength. Absorption coefficients of aerosols as a function of wavelength are calculated following Bodhaine [1995] and Weingartner et al. [2003] as

equation image

where i1 and i2 are intensities of the sample and the reference beams respectively after a sampling time interval (Δt), Q is the volume of air sampled during the time interval Δt, and A is the area of the exposed spot on the filter where aerosols are collected. C is the correction factor applied to account for any change in the absorption occurring due to aerosols on the filter over that of the airborne particles. R is an empirical correction factor and describes the change in the aethalometer response with increased particle loading on the filter. From the absorption coefficients black carbon mass concentrations are determined [Bodhaine, 1995; Weingartner et al., 2003] as

equation image

where σabs(λ) and σatn(λ) = σabs(λ)C are the mass specific absorption and attenuation cross sections respectively. The value of σatn(λ) is taken as 14625/λ nm m2 g−1, which results in an absorption efficiency of 16.6 m2 g−1 at 880 nm.

[7] The most notable uncertainties in BC mass concentration estimates using aethalometer measurements arise because of the changes in filter scattering due to aerosol loading, underestimation of the measured aethalometer signals (or BC mass concentrations) with increasing filter loads, and empirical conversion from optical absorption to BC mass [Bodhaine, 1995; Weingartner et al., 2003; Bond and Bergstrom, 2006]. In the present study wavelength-dependent values for C following Bodhaine [1995] and Bond and Bergstrom [2006] are used. The underestimation of BC mass concentrations with increasing filter loads, denoted by R, was found to be very pronounced for pure soot particles (very low SSA) while it was almost negligible for aged atmospheric aerosols [Weingartner et al., 2003]. As the annual average SSA for aerosols over Ahmedabad is found to be 0.80 ± 0.10, R is assumed to be unity, and so the uncertainties in BC mass concentrations with increasing filter loads is expected to be negligible. The optics of aethalometer allows detection of changes in light intensity of 1 in 104 which corresponds to a noise level of about 1.5 × 10−8 m−1 in absorption coefficient, which translates into 0.0015 μg m−3 of BC [Bodhaine, 1995]. The monthly mean BC mass concentrations obtained in the present study are at the least two orders of magnitude higher than the noise level BC mass concentration.

[8] BC mass concentration measured at 880 nm wavelength is considered to represent a true measure of BC in the atmosphere as at this wavelength BC is the principal absorber of light while the other aerosol components have negligible absorption at this wavelength [Bodhaine, 1995]. Hematite (Fe2O3) in dust is the other strong absorber in the atmosphere at 880 nm. However, the absorption cross section of dust is smaller than BC by more than 100 times. The absorption efficiency of 16.6 m2 g−1 has been found to satisfactorily describe the BC mass concentration levels and trends in an urban atmosphere [Hansen and Novakov, 1990; Liousse and Cachier, 1992; Ramachandran and Rajesh, 2007]. The absorption efficiency used in the study lies within the range of values (5–25 m2 g−1) [Bond and Bergstrom, 2006, and references therein]. BC mass concentration is inversely proportional to the absorption efficiency, hence, a higher value of absorption efficiency would result in lower BC mass concentrations. Thus, the BC mass concentrations reported in this study represent the minimum values. Instrumental artifacts, such as flow rate, filter spot area, and detector response, are estimated to contribute an error of ∼1% in the measured BC mass concentrations. The overall uncertainty in the BC mass concentrations reported in this study is estimated to be 10%.

[9] The number of days on which BC mass concentration measurements have been made during 2008 and the data of which are used in the study corresponds to 323 days (Table 1). BC mass concentration measurements obtained during the local festivals (Navratri and Deepavali) in 2008 are not included in the study. These two festivals are celebrated every year in India during fall season. Navratri in 2008 was celebrated between September 30 and October 9. During this festival an increase in the vehicular traffic occurs that modifies the diurnal BC mass concentration patterns [Ramachandran and Rajesh, 2007]. Deepavali was celebrated on 28 October 2008. This festival is marked by bursting firecrackers and fireworks that takes place during both the day and night for a period of 5 days, 2 days each before and after the festival. Firecrackers and fireworks can be potent sources of soot particles. Not only was the diurnal pattern modified during these festivals but also the BC mass concentrations were higher when compared to normal days over Ahmedabad [Ramachandran and Rajesh, 2007]. Owing to these reasons the number of days of BC mass concentration measurements are less than 25 days in September and October 2008, while in December 2008 measurements over Ahmedabad could be made only up to December 15, as the instrument was deployed on board a cruise that started by the end of December.

Table 1. Number of Days on Which BC Mass Concentration Measurements Were Made Over Ahmedabad During January–December 2008, the Data of Which Are Used in the Present Study
Number of days312931263130303024143016323

3.2. Aerosol Optical Properties

[10] The Moderate Resolution Imaging Spectroradiometer (MODIS) is a remote sensor onboard the two Earth Observing System (EOS) Terra and Aqua satellites [Remer et al., 2008]. Level 3 MODIS Collection V005 daily aerosol optical depth (AOD) at a 1° × 1° grid from Terra and Aqua are utilized [Remer et al., 2008]. MODIS Terra and Aqua satellites operate at an altitude of 705 km with Terra spacecraft crossing the equator at about 1030 local standard time (LST) (ascending northward) while the Aqua spacecraft crosses the equator at around 1330 LST (descending southward) [Remer et al., 2008]. MODIS Terra- and Aqua-derived aerosol products over land and oceans are tested, validated, and compared and are used extensively to investigate spatiotemporal variations in aerosol optical characteristics [e.g., Remer et al., 2008; Smirnov et al., 2006]. The predicted retrieval uncertainty of MODIS-derived AODs is found to be ±(0.05 + 0.15AOD) over land and ±(0.03 + 0.05AOD) over ocean respectively [Remer et al., 2008]. Over continental India, MODIS Terra- and Aqua-derived AODs have been validated [e.g., Ramachandran, 2007]. MODIS-derived AODs were found to compare well with in situ Aerosol Robotic Network (AERONET) Sun photometer results over Kanpur [Ramachandran, 2007]. The daily AODs obtained from Terra and Aqua corresponding to the location of Ahmedabad are used to determine the monthly mean AODs from January to December 2008.

3.3. Estimation of Aerosol Radiative Forcing

3.3.1. Necessary Inputs: Aerosol Optical Properties

[11] The radiative transfer algorithm SBDART (Santa Barbara DISORT Atmospheric Radiative Transfer) developed by Ricchiazzi et al. [1998] has been used to perform the radiative transfer calculations in the shortwave (0.2–4.0 μm) and longwave (4.0–40.0 μm) regions. SBDART computes plane-parallel radiative transfer in clear sky conditions within the Earth's atmosphere and at the surface. The model is well suited to study the radiation budget of the Earth-atmosphere system. The primary input parameters required for calculating aerosol radiative forcing are aerosol optical depth, single-scattering albedo, and asymmetry parameter (g). The Optical Properties of Aerosols and Clouds (OPAC) model developed by Hess et al. [1998] is used to retrieve the required input aerosol parameters by varying the aerosol components that contributed to the aerosol properties over Ahmedabad. The OPAC model provides optical properties of various aerosol components in the solar and terrestrial spectral wavelength range. In OPAC new mixtures can be defined from the given aerosol components to best fit the observed aerosol parameters such as optical depth, single-scattering albedo, asymmetry parameter, and so on. The aerosols are distributed in the atmosphere on the basis of scale height, which is 8 km for an urban region. Optical properties of aerosols can get modified by water uptake from the atmosphere. Taking this into account, OPAC outputs aerosol optical depth, single-scattering albedo and asymmetry parameter for aerosols which are spherical and externally mixed, at eight relative humidity (RH) (0%, 50%, 70%, 80%, 90%, 95%, 98%, and 99%) conditions. Output parameters (AOD, SSA, and g) that are crucial for aerosol radiative forcing estimation in the shortwave and longwave regions are obtained for the mean RH (Figure 1b) corresponding to each month over Ahmedabad.

[12] The urban aerosol model is made of water-soluble, insoluble, and black carbon aerosol components. The insoluble aerosol component consists mostly of soil particles with a certain amount of organic material [Hess et al., 1998]. The water-soluble aerosols originate from gas to particle conversion and are made up of various kinds of sulfates, nitrates, and organic and water-soluble substances. The water-soluble aerosol component, thus, contains more than only the sulfate aerosol that is used to describe the anthropogenic aerosol [Hess et al., 1998]. This suggests that organic carbon species are accounted for in OPAC in both insoluble and water-soluble aerosol components. The number densities of water-soluble, insoluble, and black carbon aerosols are 28,000, 1.5, and 130,000 particles cm−3 respectively [Hess et al., 1998]. In the urban aerosol model 130,000 particles cm−3 give rise to a BC mass of 7.8 μg m−3. In the present study, the monthly mean BC mass concentrations measured over Ahmedabad are used as inputs in the urban aerosol model to determine AOD, SSA, and g in the 0.2 to 40.0 μm wavelength range. The number density of BC aerosols in the urban model is varied to match the measured monthly mean BC mass concentration over Ahmedabad, while the number densities of insoluble and water-soluble aerosols are kept unchanged (refer to section 4.2 for further details).

3.3.2. Additional Inputs: Atmosphere, Ozone, Water Vapor, and Surface Albedo

[13] To perform aerosol radiative forcing calculations atmospheric profiles of temperature, pressure, columnar ozone, water vapor, and surface reflectance characteristics are necessary in addition to aerosol properties. Standard tropical atmospheric profiles of temperature and pressure are used [McClatchey et al., 1972]. Monthly mean columnar ozone and water vapor over Ahmedabad are obtained from the Total Ozone Mapping Spectrometer (TOMS) and National Center for Environmental Prediction (NCEP) reanalysis respectively. Surface albedo is important to estimate aerosol radiative forcing more accurately, especially over land regions. Surface reflectance measured by MODIS onboard Terra and Aqua satellites (8-Day, Level 3 Global 500m ISIN Grid product, MOD09A1 (Terra) and MYD09A1 (Aqua)) at seven wavelength bands centered at 0.645, 0.859, 0.469, 0.555, 1.24, 1.64, and 2.13 μm are used. Columnar ozone shows a winter low and a summer high quite consistent with the seasonal variation over the tropics (Figure 2a). Columnar water vapor is maximum over Ahmedabad during the summer monsoon months of June–July–August–September (Figure 2b). Monthly mean surface reflectance data available at the seven central wavelengths for Ahmedabad are plotted in Figure 2c. Vertical lines in Figure 2 denote the ±1σ variation from the mean value of the measured surface reflectance. The surface reflectance over Ahmedabad is higher during monsoon and premonsoon months, while it is low during winter and postmonsoon seasons. The surface reflectance data in the above wavelength bands are utilized to reproduce the spectral dependence of surface albedo in the entire shortwave and longwave regions using a combination of vegetation, sand, and water surface types [e.g., Ramachandran et al., 2006].

Figure 2.

Seasonal variation of additional inputs used in the estimation of aerosol radiative forcing over Ahmedabad. (a) Monthly mean columnar ozone (Dobson Units (DU)), (b) water vapor (cm), and (c) MODIS derived surface reflectance in seven wavelength bands. ±1σ deviation from the mean is denoted by vertical bars.

3.3.3. Calculation

[14] The aerosol radiative forcing (ARF) calculations are performed using eight radiation streams at 1 h intervals for a range of solar zenith angles and 24 h averages are obtained. Aerosol radiative forcing (ΔF) at the top of the atmosphere (TOA) and surface (SFC) is defined as the change in the net (down minus up) flux with and without aerosols and expressed as

equation image

[15] The difference between the radiative forcing at the top of the atmosphere (TOA is 100 km in this case) and the surface is designated as the atmospheric forcing (ATM) and is written as

equation image

[16] ΔFATM represents the amount of energy trapped within the atmosphere due to the presence of aerosols. If ΔFATM is positive the aerosols produce a net gain of radiative flux to the atmosphere leading to a heating (warming), while a negative ΔFATM indicates a net loss and thereby cooling.

[17] The absorption and emission processes in the longwave at different altitudes in the atmosphere, when integrated over all the wavelengths, can result in either a net gain (warming) or loss (cooling) of radiative energy [Ramaswamy, 2002], while solar radiation always warms the atmosphere (equation (4)). The infrared absorption by aerosols decrease the outgoing longwave flux while increasing the surface reaching infrared radiation resulting in a cooling of the atmosphere.

[18] ΔFATM, the amount of energy trapped in the atmosphere due to aerosols in the shortwave region, gets converted into heat. The solar heating rate can be calculated as

equation image

where ∂T/∂t is the heating rate (Kelvin per day), g is the acceleration due to gravity, Cp is the specific heat capacity of air at constant pressure, and P is the atmospheric pressure [Liou, 1980]. Large amounts of different kinds of atmospheric aerosols (e.g., water-soluble, black carbon, sea salt, and mineral dust) are concentrated from near the surface to 3 km [e.g., Intergovernmental Panel on Climate Change, 2007; Ramanathan et al., 2005] over urban, continental, and marine environments. Therefore, ΔP (in equation (5)) is considered as 300 hPa, which is equal to the pressure difference between surface and 3 km. Aerosol radiative forcing in the shortwave and longwave regions (equations (3) and (4)) are estimated over Ahmedabad. As the shortwave radiative forcing is ≥90% when compared to the longwave forcing in an absolute sense (discussed later in section 4.3), in this study only the seasonal variation of solar atmospheric heating rate (equation (5)) over Ahmedabad is calculated. The relative standard error in radiative forcing and heating rate reported here, taking into account the aerosol input parameters, uncertainties in BC mass concentrations, and flux estimates, is estimated to be 20%.

4. Results and Discussion

4.1. Monthly Mean Black Carbon Aerosol Mass Concentrations

[19] BC aerosol mass concentrations measured for 24 h/d during each month are averaged and the monthly mean BC mass concentrations are obtained (Figure 3). Vertical bars denote ±1σ from the mean, which indicates the variability in the BC mass concentrations measured during that month. BC mass concentrations in Ahmedabad exhibit a strong seasonal cycle marked by a winter high and a summer low. The seasonal mean BC mass concentrations are 11.6 ± 2.9 μg m−3 (winter (DJF)), 3.9 ± 2.5 (premonsoon (MAM)), 2.1 ± 0.8 (monsoon (JJAS)), and 10.9 ± 1.5 μg m−3 (postmonsoon (ON)) respectively.

Figure 3.

Monthly mean BC mass concentrations (μg m−3) over Ahmedabad from January to December 2008. Vertical bars denote ±1σ deviation from the mean BC mass concentration.

[20] Over an urban region no significant change in the local production of aerosols from anthropogenic activities is expected during the year [Ramachandran, 2007]. During winter the boundary layer is shallow and holds the pollutants in a smaller volume near the Earth's surface when compared to summer. During winter over Ahmedabad the winds are north-northeasterly, which bring in human-made pollutants. In addition to fossil fuel combustion in winter due to colder temperatures in Ahmedabad (Figure 1), a significant increase in the amount of open biomass burning occurs as people burn dry leaves, shrubs, paper, and waste materials to keep them warm. This leads to an increase in the amount of biomass-burning aerosols over a particular location; fossil fuel emissions and the additional biomass burning combined with the shallow boundary layer result in higher BC mass concentrations. BC mass concentrations decrease during premonsoon due to increase in the magnitude of wind speeds (Figure 1), and change in the source region and direction of wind as winds originate and travel from and through a less polluted west (arid–marine). The decrease in BC during monsoon when the rainfall is maximum can be directly related to the wet removal of aerosols by precipitation. There is less rainfall in September, resulting in an increase in BC mass (Figure 3). The wind speeds are much weaker during postmonsoon, and aerosols produced in the urban areas do not get properly ventilated or transported to downwind locations. Thus, the local sources over Ahmedabad influence the BC mass concentrations, in addition to atmospheric dynamics (boundary layer, winds, and long-range transport) and rainfall.

[21] BC mass concentrations measured over Ahmedabad are comparable to those measured in other urban locations in India, such as Hisar (1.5–7.2 μg m−3 [Ramachandran et al. 2006]), Hyderabad (3.5 μg m−3 [Moorthy et al. 2004]), and Mumbai (12.4 ± 5.1 μg m−3 [Venkataraman et al. 2002]); higher than the coastal stations Trivandrum (1.5 (summer)–5 (winter) μg m−3 [Babu and Moorthy 2002]), Goa (3 μg m−3 [Leon et al. 2001]), Visakhapatnam (0.43 (postmonsoon)–8.01 μg m−3 (winter) [Sreekanth et al., 2007]); lower than Delhi (29 ± 14 μg m−3 [Ganguly et al. 2006a]), and Kanpur (6–20 μg m−3 [Tripathi et al., 2005]).

4.2. Aerosol Optical Depth and Single-Scattering Albedo

[22] Monthly mean 0.55 μm AODs obtained from daily MODIS Terra and Aqua data are plotted in Figure 4a. AODs show a strong seasonal variation with a winter low and a summer high, which is quite consistent with earlier results [Ramachandran, 2007]. Vertical bars indicate ±1σ variation from the mean. AODs in the summer monsoon are higher (0.63) by a factor of 2 when compared to winter (0.31) over Ahmedabad. Prevailing strong convection, a deeper boundary layer providing a longer atmospheric column that is able to accommodate more number of aerosols, and hygroscopic growth of water-soluble aerosols due to higher RH (Figure 1) give rise to higher AODs during monsoon season [Ramachandran, 2007]. AODs at 0.55 μm estimated from OPAC following the procedure described in section 3.3 agree well (within ±1σ) with MODIS monthly mean AODs (Figure 4a).

Figure 4.

(a) Monthly mean MODIS AODs calculated from Terra and Aqua daily AODs at 0.55 μm in comparison with OPAC model derived AODs. Vertical bars in MODIS AODs denote the ±1σ deviation from the monthly means. See text for details. (b) Monthly mean BC mass fraction (%) over Ahmedabad, and corresponding single-scattering albedo (SSA) at 0.55 μm estimated using the measured BC mass concentrations in the OPAC model.

[23] The monthly mean BC mass concentration is highest in December (13.8 μg m−3) and lowest in June (1.6 μg m−3) over Ahmedabad (Figure 3). The number density of BC aerosols in the urban aerosol model is varied to obtain a match with the measured monthly mean BC mass concentrations over Ahmedabad to determine the aerosol optical properties (AOD, SSA, and g) (section 3.3). A number density of 229,325 particles cm−3 of BC in the urban aerosol model results in the highest monthly mean BC mass concentration measured in Ahmedabad, while 26,165 BC particles correspond to the lowest monthly mean BC mass. The mean number density of BC aerosols increases by about an order of magnitude between July (lowest) and January (highest), which is quite consistent with the variation observed in the monthly mean BC mass (Figure 3).

[24] SSA at 0.55 μm derived from the OPAC model using the measured BC aerosol mass concentration as an input varies from 0.64 in December to 0.95 in July over Ahmedabad. Seasonal mean SSAs over Ahmedabad are estimated as 0.68 ± 0.04 (DJF, winter), 0.83 ± 0.08 (MAM, premonsoon), 0.93 ± 0.03 (JJAS, monsoon), and 0.69 ± 0.04 (ON, postmonsoon) respectively. The mass fraction of the water-soluble aerosol component increases as RH increases due to its hygroscopic nature. However, in order to maintain a constant mass mixing ratio (100%), an inherent property of the model, the mass fractions of insoluble and BC aerosols decrease in urban aerosols (Table 2, Figure 4b) resulting in higher SSA, as seen during June–August over Ahmedabad. SSAs obtained during winter and postmonsoon over Ahmedabad are lower than those measured over south Asia and India, for example, Hanimaadhoo (6.8°N, 73.2°E) (0.9 (winter)–1.0 (monsoon [Ramana and Ramanathan, 2006]), six provinces in China (0.851–0.802, mean 0.832 [Qui et al., 2004]), Pune (0.81 [Pandithurai et al., 2004]), Kanpur (0.76 [Tripathi et al., 2005]), Hisar (0.76–0.88 [Ramachandran et al., 2006]), and comparable to that of Delhi (0.68 [Ganguly et al., 2006a]). High BC mass and lower SSA over an urban location can have significant consequences on the radiation budget as BC can absorb both incoming solar and outgoing terrestrial longwave radiation.

Table 2. Monthly Mean Relative Humidity Over Ahmedabad During 2008a
MonthsRH (%)0% RH50% RH90% RHAlbedo (0.555 μm)
  • a

    Monthly mean variation in aerosol mass fraction (%) of insoluble (IS), water-soluble (WS) and black carbon (BC) for 0, 50, and 90% RH. Monthly mean surface albedo from MODIS Terra and Aqua at the central wavelength of 0.555 μm (0.545 to 0.565 μm band) over Ahmedabad from January to December 2008. RH is relative humidity.


4.3. Shortwave and Longwave Aerosol Radiative Forcing and Heating Rate

[25] Figure 5 shows the monthly mean shortwave ARF for Ahmedabad at TOA, SFC, and ATM (see section 3.3, equations (3) and (4)) as a function of RH. The figure shows that SFC ARF decreases during the summer, while TOA ARF increases. For both SFC and TOA, increase in RH corresponds to a more negative ARF. The radiative forcing at TOA changes sign over higher reflectance surfaces (e.g., land, snow) when absorbing aerosols are abundant [Intergovernmental Panel on Climate Change, 2007; Ramanathan et al., 2005]. Thus TOA forcing becomes positive during October–March when BC aerosols are abundant. During summer monsoon, although the surface reflectance is higher (Figure 2) as the BC mass concentrations are low (Figure 3), TOA forcing is negative. ATM warming is more positive during winter when compared to summer. ATM warming exhibits the least variation when RH increases (Figure 5). The magnitude of ATM warming does not change significantly as RH increases because the variation in SFC forcing is compensated by the decrease or increase and sign change of TOA forcing (Figure 5).

Figure 5.

Clear sky shortwave (0.2–4.0 μm) aerosol radiative forcing (a) at the top of the atmosphere, (b) in the atmosphere, and (c) at the surface corresponding to 0% relative humidity (RH), 50% RH, and 90% RH.

[26] The objective of the present study is to estimate the net ARF at TOA, SFC, and ATM. The inclusion of aerosol vertical profiles is not expected to significantly modify the net ARF, heating rate, overall outcome, and conclusions. However, in this study the influence of aerosol vertical profile on ARF and heating rate on a seasonal mean basis over an urban region (Ahmedabad) is estimated to verify the same. As simultaneous lidar measured aerosol profiles are not available in 2008 for Ahmedabad, seasonal model aerosol vertical profiles similar to those obtained during 2002–2005 using micropulse lidar [Ganguly et al., 2006b] are constructed and utilized. By doing so the vertical structure of aerosols similar to those measured earlier over Ahmedabad is maintained, while AOD corresponds to 2008. The year-to-year intraseasonal variations in aerosol vertical profiles measured over Ahmedabad are found to be much less [Ganguly et al., 2006b]. Therefore, uncertainty, if any, due to the noninclusion of yearly variation in aerosol vertical profiles is expected to be small and is not accounted for in the present study. The seasonal mean aerosol extinction profiles obtained at the lidar wavelength of 0.523 μm during winter (DJF), premonsoon (MAM), monsoon (JJAS), and postmonsoon (ON) have been scaled by the seasonal mean AODs (Figure 4a). Seasonal mean aerosol input parameters (AOD, SSA, and g) are computed using the seasonal mean BC mass concentrations following the procedure described in section 3.3. Seasonal mean shortwave ARF and solar heating rates obtained over Ahmedabad by including the measured aerosol vertical profile are compared with those obtained by not including the aerosol vertical profiles in Figure 6. The inclusion of vertical profiles of aerosols had no effect on ARF at TOA, SFC, ATM, and the heating rates as expected (Figure 6).

Figure 6.

Shortwave aerosol radiative forcing calculated for different seasons with and without using the aerosol vertical profile information over Ahmedabad corresponding to (a) winter, (b) premonsoon, (c) monsoon, and (d) postmonsoon seasons. Heating rates (Kelvin per day) as function of altitude for (e) winter, (f) premonsoon, (g) monsoon, and (h) postmonsoon. (i) Aerosol radiative forcing at the top of the atmosphere, surface, and in the atmosphere for different seasons obtained with and without aerosol vertical profiles. Heating rates (Kelvin per day) are given in brackets.

[27] Longwave forcings at SFC show a strong seasonal and RH dependence (Figure 7). In an absolute sense, longwave SFC forcing is ≤10% when compared to shortwave SFC forcing (Figure 5). TOA forcing in longwave also becomes more positive when RH increases. The magnitude of longwave ATM forcing decreases from January to June. Longwave ATM cooling is lowest in the month of June and is less than −0.1 W m−2 at all RH values (Figure 7). During July–August longwave ATM forcing becomes slightly positive (<0.5 W m−2) due to the increase in columnar water vapor. Over Ahmedabad the columnar water vapor increases by a factor of 3–4 during monsoon when compared to winter (Figure 1). It is seen that shortwave TOA forcing can become positive when abundant absorbing aerosols are found over higher reflectance regions, thereby resulting in a more positive ATM warming; while longwave ATM cooling changes to warming when water vapor increases. Thus, when higher amount of BC aerosols and water vapor are present over high reflectance surfaces the net (shortwave + longwave) ATM warming becomes more positive, while when water vapor amount is small the longwave ATM cooling partially counterbalances the shortwave ATM warming.

Figure 7.

Longwave (4.0–40.0 μm) aerosol radiative forcing (a) at the top of the atmosphere, (b) in the atmosphere, and (c) at the surface corresponding to 0, 50, and 90% RH.

[28] To delineate the impact of BC on the Earth–atmosphere radiation budget over urban regions, an aerosol model without including BC aerosols is constructed and the radiative effects are investigated. Aerosol input parameters (AOD, SSA, and g) are computed in the shortwave and longwave regions following the procedure described in section 3.3 for an urban aerosol model having no BC aerosols for a range of RH varying from 0 to 90%. The estimated ARF in the shortwave and longwave and the solar heating rates are plotted in Figure 8 from 0 to 90% RH. TOA and SFC forcing in the shortwave region increase as RH increases, resulting in an almost identical ATM warming. TOA forcing is negative for all RH, in contrast to the variation seen in TOA forcing when BC aerosols are present (Figure 5). SFC forcing is less negative when BC aerosols are absent when compared to the SFC forcing obtained including BC aerosols (Figure 5). ATM warms by about 7 W m−2 when no BC aerosols are present, which translates into a heating rate of ∼0.21 K/d. In the absence of BC aerosols ATM warming is less positive; this small warming occurs because of the absorption of solar radiation by water-soluble and insoluble aerosols. SSA estimated using OPAC model for zero BC is found to vary less as RH increases when compared to those obtained including BC aerosols (Table 3). Longwave TOA, SFC, and ATM in the zero BC case are on an average about 75% of the respective forcings obtained with BC aerosols (Figure 8b), suggesting that BC aerosols contribute about 25% to longwave forcing. These results indicate that on an average over an urban region BC aerosols can contribute about 60% to shortwave ATM warming and 25% to longwave ATM forcing.

Figure 8.

Aerosol radiative forcing at the top of the atmosphere, surface and in the atmosphere corresponding to zero BC mass concentrations as function of RH varying from 0 to 90% in the (a) shortwave and (b) longwave regions. (c) Heating rates (Kelvin per day) for zero BC mass concentrations in the 0 to 90% RH range.

Table 3. Single-Scattering Albedo at 0.55 μm Estimated Using OPAC Model as a Function of Relative Humidity for January With and Without BC Aerosols
RH (%)aSingle-Scattering Albedo
With BCZero BC
  • a

    RH is Relative Humidity.


[29] The monthly mean solar heating rates calculated from equation (5) for the shortwave ATM forcing shown in Figure 5 and averaged from 0 to 90% RH are plotted in Figure 9. The heating rates are higher during postmonsoon and winter. The heating rates are higher than 1.6 K/d during October–December and highest in December when BC mass concentration is also highest (∼14 μg m−3) (Figure 3). The seasonal mean heating rates are estimated to be 1.53 K/d (winter), 0.93 K/d (premonsoon), 0.65 K/d (monsoon), and 1.70 K/d (postmonsoon). The heating rates are higher than factors of 7 (winter), 4 (premonsoon), 3 (monsoon), and 8 (postmonsoon) respectively when compared to the mean heating rate without BC (∼0.21 K/d), thus highlighting the significant influence BC aerosols have on radiation budget and climate impact. The results from the present study gain further importance as not only the aerosol abundances are increasing over India in particular and south-southeast Asia in general, but also their sources exhibit large spatial and temporal variations [e.g., Qiu et al., 2004; Ramanathan et al., 2005; Ramachandran, 2007]. The large positive atmospheric warming and negative surface forcings estimated in the present study can have a significant influence on atmospheric stability and cloud formation in the tropics and hence can lead to a weaker hydrological cycle [Ramanathan et al., 2005].

Figure 9.

Monthly mean heating rates (Kelvin per day) corresponding to the monthly mean shortwave aerosol atmosphere warming shown in Figure 5. See text for details.

5. Summary and Conclusions

[30] BC aerosol mass concentrations measured in Ahmedabad, an urban location in western India, are analyzed. The radiative effects of aerosols have been estimated utilizing the measured BC mass concentrations and aerosol optical depths. Aerosol radiative forcings so obtained are contrasted with forcings estimated without BC to assess the influence of BC on the Earth–atmosphere radiation budget and climate.

[31] The major conclusions of the study can be summarized as follows.

[32] 1. BC mass concentrations over Ahmedabad vary from a low of 2 μg m−3 during summer monsoon to a high of 11 μg m−3 during winter and the postmonsoon season. AOD at 0.55 μm during the summer monsoon is a factor of 2 higher (0.63) than that in winter (0.31) in Ahmedabad.

[33] 2. BC mass concentrations over Ahmedabad are influenced by local sources and meteorology (boundary layer, rainfall, winds, and long-range transport). Lower BC mass concentrations during monsoon occur due to wet removal of BC aerosols near the surface, while higher BC mass concentrations during winter and postmonsoon arise due to shallow boundary layer, long-range transport, and low wind speeds. BC mass concentrations decrease from winter to premonsoon due to variation in the source regions and pathways of long-range transport as the winds change direction from north (polluted) to west (arid–marine).

[34] 3. SSA estimated using the measured black carbon aerosol mass concentration as input in an aerosol optical properties model over Ahmedabad is found to be 0.68 (winter), 0.83 (premonsoon), 0.93 (monsoon), and 0.69 (postmonsoon) respectively.

[35] 4. Shortwave surface forcing is maximum in December (−60 W m−2) and minimum in July (−23 W m−2). SFC forcing becomes more negative when RH increases. TOA forcing becomes positive when BC amount increases and RH is low. The magnitude of ATM warming does not show significant variation with RH as the more negative (less negative) SFC forcing is compensated by the more negative (less negative or positive) TOA forcing. ATM warming is about 50 W m−2 during winter and increases to 60 W m−2 during postmonsoon.

[36] 5. Positive TOA ARF during postmonsoon and winter becomes negative during premonsoon and monsoon because of low BC though the surface reflectance is higher. ATM forcing is more positive during postmonsoon and winter in spite of lower AODs due to lower SSAs. In contrast, ATM forcing is less positive during premonsoon and monsoon regardless of higher AODs due to higher SSAs.

[37] 6. Forcing estimates confirm that inclusion or noninclusion of aerosol vertical profiles does not significantly modify the net top of the atmosphere, surface, and atmosphere aerosol radiative forcing and the heating rates over an urban region.

[38] 7. Longwave forcings at SFC are <3 W m−2 during January–December at 0% RH. Longwave forcing is about 10% of shortwave forcing. Longwave ATM cooling is <−2 W m−2 at 0% RH during all the seasons except monsoon. The magnitude of longwave ATM forcing decreases from January to June. During July–August longwave warms the atmosphere due to increase in columnar water vapor. Shortwave ATM warming becomes more positive when absorbing aerosols increase over higher albedo surfaces, while longwave warms the atmosphere when water vapor increases. Thus, when the amounts of BC and water vapor are high over continental regions, the net (shortwave + longwave) ATM warming will be higher.

[39] 8. Shortwave ARF TOA forcing is negative for all RH in the absence of BC aerosols. Shortwave ATM warming is only about 15% when no BC aerosols are present in the atmosphere. The heating rates are small (∼0.21 K/d) when BC aerosols are absent. Results suggest that on an average over an urban region BC aerosols alone can contribute about 60 and 25% of shortwave and longwave ATM forcing respectively.

[40] 9. The seasonal mean heating rates are higher than 1.5 K/d in winter and postmonsoon, peaking in December at 1.76 K/d. The heating rates are a factor of 7 higher during postmonsoon and winter when compared to the heating rate obtained without BC aerosols in the atmosphere. Even during premonsoon and monsoon seasons the heating rates are 3 times higher when compared to a zero BC atmosphere, emphasizing the significant influence BC aerosols exert on radiation budget and climate.

[41] Considering the potential radiative and climate impacts of black carbon (higher atmospheric warming and evaporation of low level clouds) combined with comparatively higher black carbon mass concentrations and their increasing abundances, absorbing aerosols could well be a causative agent contributing to the decreasing trend in monsoon rainfall over different regions in India, which, however, needs to be examined in detail.


[42] We thank ISRO-GBP, ISRO Headquarters, Bengaluru, for partial funding support. MODIS AODs and TOMS ozone are downloaded from the GES-DISC, NASA. Winds are downloaded from Ahmedabad station rainfall is obtained from NOAA, NESDIS National Climatic Data Center, USA. Thanks are due to T. A. Rajesh for the prompt maintenance of the aethalometer and to Rohit Srivastava for his help in calculating the vertical profiles of aerosol radiative forcing.