Spatiotemporal heterogeneity in aerosol radiative forcing and heating rate have been studied over Bay of Bengal and Arabian Sea during premonsoon (March–May 2006) using aerosol optical depth (AOD), total mass, aerosol chemical composition, and radiative transfer model. Mean 0.5 μm AOD over Bay of Bengal and Arabian Sea is 0.36 and 0.25, respectively. Water-soluble aerosols, sea salt, and mineral dust constitute ∼98% of total aerosol mass while black carbon aerosols contribute ≤2% over the two oceanic regions. Sensitivity tests reveal that (1) curvature effect in AOD spectra has insignificant impact in modifying the aerosol radiative forcing and heating rate and (2) the net Earth-atmosphere energy content shows minor differences when aerosol vertical profiles are used. Over Bay of Bengal the average aerosol forcing is estimated to be −12.0, −22.4, and 10.4 W m−2 at the top of the atmosphere (TOA), at the surface (SFC), and in atmosphere (ATM), respectively. The average aerosol radiative forcing is less negative over Arabian Sea and is −10.5, −15.8, and 5.3 W m−2 at TOA, SFC, and ATM, respectively. Aerosol radiative forcing decreases in magnitude from north to south over Bay of Bengal whereas an opposite trend is noteworthy over Arabian Sea. The average atmospheric heating rate over Bay of Bengal is ∼0.3 K/d, a factor of 2 higher than that over Arabian Sea. Furthermore, ATM warming and associated heating rate are the lowest compared to earlier results as scattering aerosols are dominant during premonsoon (March–May). These results have implications to the assessment of regional and seasonal climate impacts.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Atmospheric aerosols from both natural and anthropogenic sources affect the Earth-atmosphere radiation budget directly by scattering and absorbing the incoming solar radiation, and indirectly by modifying the cloud radiative properties through altering the cloud microphysical properties. Because of the short residence times and diverse aerosol types, the direct and indirect effects of aerosols exhibit large spatial and temporal variations. The direct and indirect aerosol radiative forcings remain a significant uncertainty for climate studies [Intergovernmental Panel on Climate Change (IPCC), 2007]. The potential for aerosol forcing of climate can vary according to regional differences in aerosol columnar concentration as well as its chemical composition [Eldering et al., 2002]. It is also well known that different aerosols interact with radiation in different ways; for example black carbon is highly absorbing and has a warming effect while sulfate is highly scattering and exhibit cooling effect in the atmosphere. Thus, knowledge of aerosol chemical composition is important to determine the scattering and absorption characteristics of aerosols. Aerosols over the marine regions comprise mainly sea salt particles, and mineral dust transported from arid and semiarid regions of surrounding landmasses in addition to anthropogenically produced aerosol particles. On a global, annual mean scale natural aerosols dominate contributing about 55–60% of aerosol emissions and columnar aerosol optical depths [IPCC, 2007]. This scenario would however be different downwind of major source regions where anthropogenic aerosols dominate the aerosol emissions and optical depths [e.g., Ramanathan et al., 2001].
 An accurate determination of aerosol radiative forcing is essential to estimate better the climate impact of aerosols on regional and global scales. Recently, an Integrated Campaign for Aerosols, gases and Radiation Budget (ICARB) was conducted during the premonsoon or intermonsoon season of March–May 2006. ICARB, a multiplatform and a multi-institutional campaign, was conducted with an aim to capture the regional and temporal variability in aerosol sources and sinks, natural and anthropogenic aerosol hot spots, and their radiative impacts over the Bay of Bengal, Arabian Sea and India through intensive simultaneous measurements [Moorthy et al., 2008]. ICARB campaign consisted of ocean, land and air segments. ICARB ocean segment comprised two cruises, each of about a month duration over the Bay of Bengal and the Arabian Sea (Figure 1). Measurements of optical, physical and chemical characteristics of aerosols were made on board the cruises [e.g., Kedia and Ramachandran, 2008; Moorthy et al., 2008; Nair et al., 2008; Kumar et al., 2008].
 Aerosol optical depth is directly proportional to aerosol loading and the size distribution of aerosol mass burden in the atmospheric column; typically in an aerosol size distribution submicron aerosols will be orders of magnitude higher than supermicron particles. The size distribution is crucial to determine the single scattering albedo (SSA), as the value of SSA (whether high or low) is determined by the ratio of the number of absorbing to scattering particles in a size distribution. Asymmetry parameter depends both on the size distribution and chemical composition of aerosols. Single scattering albedo and asymmetry parameter also vary as a function of relative humidity. Thus, optical (aerosol optical depths), physical (mass concentration and/or size distribution), and chemical (composition) characteristics of aerosols are necessary to determine single scattering albedo, asymmetry parameter, and hence, aerosol radiative forcings. In this study we report aerosol direct radiative forcings estimated using a discrete ordinate radiative transfer model in which simultaneously measured aerosol optical depths (AODs), mass concentration and chemical composition of total suspended particulate mass (TSP) in the marine atmospheric boundary layer of the Bay of Bengal and Arabian Sea are used as inputs.
2. Cruise Track, Meteorological Conditions, and Wind Patterns
 The ocean segment of ICARB was conducted on board ORV Sagar Kanya during March–May, 2006. The ICARB cruise campaign was divided into two phases. The first phase of ICARB was conducted over the Bay of Bengal between 18 March and 12 April 2006 (Figure 1a) and the second phase of the cruise was undertaken in the Arabian Sea from 18 April to 10 May 2006 (Figure 1b). The mean surface level wind speed was found to be 4.5 ± 1.8 m s−1 over the Bay of Bengal while over the Arabian Sea the wind speed was 5.4 ± 2.1 m s−1 [Kedia and Ramachandran, 2008]. The daily mean wind speeds were found to vary from ∼2 m s−1 on 20 March to a high of ∼8 m s−1 on 11 April over the Bay of Bengal. Over the Arabian Sea on an average the wind speeds were higher than the Bay of Bengal region. The wind speeds were higher than 5 m s−1 during 26–30 April over the Arabian Sea. The mean relative humidity was found to be around 73 ± 6% (Bay of Bengal) and 72 ± 3% (Arabian Sea), and the daily mean RH was found to vary by about 5–10% from the mean. The sky conditions were generally clear during the cruises over the Bay of Bengal and the Arabian Sea, while on some days occasionally cloud patches were observed.
 The back trajectory analysis provides a three dimensional (latitude, longitude, and height) description of the pathways followed by air parcels as a function of time. Air mass back trajectories at different heights are important to identify the source regions and the transport pathways of the pollutants before they reach the measurement location when analyzing the columnar properties of atmospheric constituents. Seven day air back trajectory analysis has been performed considering the residence time of different types of aerosols, which is about a week in the lower atmosphere. Air back trajectories are calculated for each day corresponding to 1200 Indian Standard Time (+0530 GMT) and for the mean latitude-longitude position of the ship (Figures 2a–2f) at different heights using the HYSPLIT meteorological model's vertical velocity fields [Draxler and Hess, 1998]. The 10 m air back trajectories are quite close to the surface and aerosols at this height will settle faster; in addition, the 10 m air back trajectories are more or less similar to the air back trajectories corresponding to 100 m. Thus, due to the above reasons, and for the purposes of clarity and illustration, air back trajectory analysis is restricted to those obtained at 100, 1000 and 2500 m heights. Air back trajectories over the Bay of Bengal and Arabian Sea originate from different arid/semiarid, continental and marine locations suggesting different source regions and aerosol types (Figures 2a–2f). The air back trajectories (Figures 2a–2c) originate from arid/semiarid regions (Pakistan, Iran and Saudi Arabia) and pass through continental India and Indo-Gangetic plain before reaching the Bay of Bengal. In contrast over the Arabian Sea most of the air back trajectories are of marine origin (Bay of Bengal, Arabian Sea, Figures 2d–2f), though on a few days back trajectories originate and pass through continental India before reaching the Arabian Sea.
 An indigenously developed hand-held Sun photometer was used for the measurement of aerosol optical depths at five wavelength bands centered around 0.40, 0.50, 0.65, 0.75, and 0.875 μm. The bandwidths (FWHM) of the filters are about 0.01 μm and the total field of view of the instrument is 8° [Kedia and Ramachandran, 2008]. The hand-held Sun photometer has been used successfully in many studies including INDOEX [Ramanathan et al., 2001]. The Sun photometer consists of an interference filter, photodiode and necessary electronics. Sun photometer observations were taken in a moving frame over the ocean, with the ship moving at an average speed of about 20 km/h. The ship was sailing a distance of more than 400 km in a day. About 40 observations were made each day at 15 min interval in all the wavelength bands from 0800 Local Standard Time (LST) to 1700 LST at different solar zenith angles during clear sky conditions [Kedia and Ramachandran, 2008]. AOD measurements were conducted on 26 days over the Bay of Bengal and 23 days over the Arabian Sea.
 The Sun photometer is periodically calibrated and Io, the solar radiation intensities for zero airmass are obtained using Langley plot technique for all the wavelengths from the measurements conducted at Gurushikhar, Mount Abu (24.6°N, 72.7°E), a hill station and a relatively clean site located at a height of about 1.7 km above MSL. These Io values are used in Beer-Lambert's law to derive the optical depths. The uncertainties in the optical depth measurements arise from (1) instrumental error due to bias and precision and (2) ignoring the forward scattering contribution to the measured irradiance. The Sun photometer is manually aimed at the Sun with the help of a Sun guider and the peak intensity is recorded during clear sky conditions. The Sun guider also helps against errors in measurements due to pitching and/or rolling of the ship; however, as the cruise campaign was conducted during premonsoon season the pitching and/or rolling effects were minimal. The solar radiation intensities are measured with an accuracy better than 1%. The forward scattered radiation within the field of the view of the photometer is found to decrease by <8% at 0.4 μm and 5% at 0.875 μm. These results are obtained based on observations made during clear sky conditions over Gurushikhar in Mount Abu where the Sun photometer is calibrated. The aerosol optical depth at any given wavelength is obtained after subtracting the contribution from Rayleigh scattering (scattering due to air molecules), ozone, water vapor etc. Typically, at 0.4 μm Rayleigh scattering contributes about 40% and aerosols contribute the rest to the total optical depth; at 0.875 μm the contribution from Rayleigh scattering decreases to 5% while the contribution due to aerosols increases to >90%. The maximum uncertainty in retrieved aerosol optical depths due to errors in measurements and assumptions involved is estimated to be <15% [Kedia and Ramachandran, 2008]. The coordinates of the ship (latitude, longitude and altitude) were measured using a custom made GPS at 1-s resolution with an accuracy better than 1/100th of a minute for latitude-longitude, and 6–10 m for altitude. Since daily mean AOD spectra are only used and as the spatial coverage is large in a day, any uncertainty arising due to the ship position will be negligible and is not accounted for in this study.
3.2. Physical and Chemical Characteristics: High-Volume Aerosol Sampling
 Bulk-aerosol samples were collected on board using PALLFLEX™ tissuquartz filters (20 × 25 cm2) by operating a high-volume sampler (HVS) at a flow rate of about 1.5 m3 min−1 over a time period ranging from 15 to 22 h [Kumar et al., 2008]. The tissuquartz filter has 99.9% collection efficiency for particles of size 0.3 μm, and the maximum cutoff radius is about 10 μm [Ramachandran et al., 2006]. Although particles of size <0.3 μm are retained on the filter, the percentage contribution of lower size particles to total suspended particulate (TSP) mass will be quite small. For example, the mass per particle decreases by 2–3 orders of magnitude when the mode radius of the particle decreases from 0.3 to 0.03 μm [Hess et al., 1998]. In a given volume of aerosols the number of large particles is orders of magnitude less compared to that of small particles, and hence, particles of size ≥10 μm have relatively insignificant contribution to aerosol optical depth. It is found that the optically effective aerosol radius range of 0.05 μm to 10 μm contributes the maximum to aerosol optical depth, as Mie scattering contribution of smaller particles (<0.05 μm) is only marginal [d'Almeida et al., 1991]. Though aerosols smaller than 0.1 μm are abundant in the atmosphere their residence times are shorter as they get converted into accumulation mode particles through gas to particle conversion mechanism. In addition, their contribution to total mass will be less. While aerosols larger than 10 μm can contribute to total mass, their role in optical properties and radiative transfer are however limited, because of lower number density and shorter residence times [d'Almeida et al., 1991].
 The flow rate of HVS calibrated before and after the cruise, was stable within 5%. Subsequently, aerosol samples were analyzed in the laboratory for their physical and chemical characterization. The TSP mass concentrations were obtained gravimetrically by weighing the full filters before and after sampling. All samples were equilibrated at a relative humidity of 50 ± 5% and at 22 ± 1°C for 5–6 h prior to their weighing [Kumar et al., 2008]. The error in the measured TSP calculated by repeated measurements of filter weight is estimated to be about 15%. The chemical composition of aerosols was obtained by analyzing water-soluble (NH4+, K+, Mg2+, Ca2+, Cl−, NO3−, SO42−), crustal element (Al, Fe, Ca) and carbonaceous aerosols [Kumar et al., 2008]. Elemental carbon (EC) and organic carbon (OC) mass concentrations are determined using EC-OC analyzer [Kumar et al., 2008]. The abundance of Aluminum (Al) is generally used as an indicator of mineral dust aerosols [Duce and Tindale, 1991]. In this study, concentration of mineral dust was estimated using Al as proxy, based on the assumption that the ratio of Al in minerals to be the same as in the upper continental crust with Al content of 8.04%. The linear regression analysis of Fe and Ca with Al yielded significant regression coefficients indicating that there was no fractionation of mineral dust over the Bay of Bengal and Arabian Sea [Kumar et al., 2008].
3.3. Aerosol Optical Properties: OPAC Model
 The principal input parameters required for calculating aerosol radiative forcing are aerosol optical depth, single scattering albedo (SSA) and asymmetry parameter (g). Optical Properties of Aerosols and Clouds (OPAC) model developed by Hess et al.  is used to reconstruct the required input aerosol parameters by varying the aerosol components that contributed to the aerosol properties over the Bay of Bengal and Arabian Sea. OPAC outputs aerosol optical depth, single scattering albedo and asymmetry parameter assuming the aerosols as externally mixed and spherical, at eight relative humidity (RH) (0%, 50%, 70%, 80%, 90%, 95%, 98% and 99%) conditions in the 0.25–4.0 μm wavelength region. Among the ten aerosol components in OPAC which include insoluble and sulfate droplets, the most suitable aerosol components based on the aerosol source regions and transport pathways over the Bay of Bengal and Arabian Sea (Figures 2a–2f) can be water-soluble, black carbon (or elemental carbon), sea salt, and mineral dust. The number concentrations of these aerosol components are altered to match the daily mean measured AOD spectra, TSP mass and the mass concentrations of chemical constituents until the following conditions are satisfied: (1) the root mean square (rms) difference between the measured and model AOD spectra is <0.03, thus, constraining the rms difference to within 0.10 AOD, (2) the OPAC estimated total mass concentrations at 50% RH lie within ±1σ of the HVS measured TSP mass, (3) OPAC estimated mass concentrations of water-soluble, black carbon, sea salt and mineral dust should be within ±1σ of the respective concentrations analyzed from HVS at 50% RH [Ramachandran et al., 2006], and (4) Ångström wavelength exponent α determined for the measured AODs in the 0.4–0.875 μm wavelength region are consistent with model derived α values. The mass concentrations of nitrate (NO3−), sulfate (SO42−), magnesium (Mg2+), potassium (K+), calcium (Ca2+), and organic carbon from HVS are summed to obtain the mass of water-soluble aerosols, and compared with the water-soluble aerosol mass concentration from OPAC. The water-soluble aerosols in OPAC correspond to aerosols that originate from gas to particle conversion mechanism and consist of various kinds of sulfates, nitrates, organics, and water-soluble substances. Elemental carbon (also known as black carbon) from HVS is compared with black carbon obtained from OPAC. Sea salt concentration from HVS is calculated from the mass concentrations of sodium (Na+) and chloride (Cl−) using the relation [Na+] × 1.47 + [Cl−] (μg m−3). Note that water-soluble aerosols and sea salt are treated separately in this study. Mineral dust calculated using Al concentration as proxy is compared with mineral dust obtained from model.
3.4. Aerosol Radiative Forcing Calculation
 The radiative transfer calculations are made using the Santa Barbara Discrete Ordinate Radiative Transfer (SBDART) [Ricchiazzi et al., 1998] algorithm which is proven to be an useful tool to address issues related to Earth-atmosphere radiation budget. SBDART computes plane-parallel radiative transfer in clear sky conditions within the Earth's atmosphere and at the surface. Aerosol optical depth, single scattering albedo and asymmetry parameter determined in the wavelength range of 0.25–4.0 μm following the procedure described in section 3.3 are used as inputs in SBDART for radiative transfer calculations. The shortwave aerosol radiative forcing (ARF) calculations are performed using 8 radiation streams at 1 h interval 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) can be defined as the change between the net (down minus up) flux with and without aerosols as
 The difference between the radiative forcing at the top of the atmosphere (which is 100 km in this case) and the surface is defined as the atmospheric forcing (ATM) and can be written as
 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.
 The radiative and subsequently the climate implications of aerosols are assessed in terms of the atmospheric heating rate. The atmospheric forcing in W m−2 (equation (2)) indicates the amount of radiative flux (energy) absorbed by the aerosols. This energy which is converted into heat is calculated as heating rate in K/d following equation (3).
where ∂T/∂t is the heating rate (K/d), 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 amount of different kinds of atmospheric aerosols (water-soluble, black carbon, sea salt and mineral dust) are concentrated from near surface to up to 3 km [e.g., IPCC, 2007; Ramanathan et al., 2001] over urban, continental and marine environments. Therefore, ΔP (in equation (3)) is considered as 300 hPa which is equal to the pressure difference between surface and 3 km.
 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]. The columnar ozone and water vapor for tropical atmosphere are 253 Dobson Units and 4.12 cm, respectively [Ricchiazzi et al., 1998]. The shortwave clear sky aerosol radiative forcings did not exhibit any significant differences for varying water vapor and ozone column amounts in the tropical atmosphere [Ramachandran et al., 2006]. Surface reflectance is a crucial parameter which can introduce errors in the aerosol radiative forcing estimates [Wielicki et al., 2005]. Surface reflectance measured by MODIS on board 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. The MODIS derived surface albedo corresponding to the seven central wavelengths are compared with the surface albedo for sea water [Viollier, 1980] as given in SBDART (Table 1). The ±1σ variation from the mean MODIS albedo over the Bay of Bengal and Arabian Sea during March–April–May is also given in Table 1. The sea water surface albedo values are higher than the MODIS derived value at 0.555 and 0.645 μm, while it is lower at 0.859 μm (Table 1); the surface albedo for sea water is 0 at the other wavelengths. For the rest of the wavelengths where MODIS surface albedo values are not available, sea water surface reflectance characteristics [Viollier, 1980] as given in SBDART are used. As the surface reflectance over marine region is lower than land, the aerosol radiative forcing is found to differ only by less than 0.5% when sea water albedo is used instead of MODIS measured albedo. The relative standard error in radiative forcing reported in the study, taking into account the uncertainties in aerosol input parameters and flux estimates, is found to be 20%.
Table 1. MODIS Terra/Aqua Derived Surface Reflectance During March–May 2006 in Comparison With SBDART Albedo Values for Seawater
Bay of Bengal
4. Results and Discussion
4.1. Aerosol Optical Depths, Total Aerosol Mass, and Chemical Species
 The day to day variation in 0.5 μm aerosol optical depths measured in situ using Sun photometer is plotted in Figure 3a. AODs are about 0.4 in the initial phase of the cruise which increase to >0.7 during 23–24 March. AODs decrease to reach values of 0.3 during 25–30 March after which the AODs increase over the Bay of Bengal. AODs show an increasing trend during 7–13 April when the ship was sailing toward Sri Lanka and then India (Figure 1a). AODs measured over the Arabian Sea during ICARB are lower when compared to the Bay of Bengal. AODs over the Arabian Sea are higher than 0.5 during 25–27 April. AODs are low during the first week of May. AODs increase to ∼0.3 thereafter. High AODs occur when the ship was cruising near the coast, and when the winds were found to originate and pass through arid and urban regions before reaching the measurement locations (Figure 2). The spatial and temporal variability in AODs across the Bay of Bengal and Arabian Sea could be due to aerosols produced locally and long-range transport (Figure 2). OPAC model estimated aerosol optical depths following the procedure described in section 3.3 are found to agree well with the measured AODs (Figure 3a).
 Aerosol mass concentration obtained from HVS during the cruise period corresponds to 50 ± 5% RH, while the OPAC estimated mass concentrations are available for different RH values ranging from 0% to 99%. In the present study we compare the total mass and species concentrations obtained from OPAC at 50% RH with values obtained from HVS at a mean RH of 50%. The total mass concentrations estimated from OPAC and measured using HVS are found to agree very well (Figure 3b). The total suspended particulate mass was in the range of 5–47 μg m−3 over the Bay of Bengal and Arabian Sea, and exhibit spatiotemporal variability similar to AODs. TSP is found to be 35–40 μg m−3 when the ship started from Chennai coast and sailed toward the northern Bay of Bengal. TSP is higher than 40 μg m−3 when the ship was cruising near the densely polluted Indo-Gangetic plain and Kolkata. The highest value of TSP occurs on 25 March (47 μg m−3) when the ship was in mid Bay of Bengal (Figure 1a). TSP starts decreasing and reaches a minimum value of 5 μg m−3 on 7th April when the AOD is also minimum. TSP started increasing when the ship was cruising near Sri Lanka and moving toward Kochi. TSP is high on 26 April (45 μg m−3) over the Arabian Sea when the ship was near the coastal Indian region. A decreasing trend in TSP is observed when the ship was in the mid Arabian Sea (≤20 μg m−3) (Figure 1b). The highest TSP mass > 45 μg m−3 was seen on 8 May when the ship was over the northern Arabian Sea near west coast. OPAC estimated mass concentrations are slightly lower than the HVS mass concentrations. OPAC mass concentrations are lower because in OPAC aerosol particles up to 7.5 μm radius only are considered for calculation of total mass [Hess et al., 1998], whereas in HVS the upper cutoff radius is about 10 μm [Ramachandran et al., 2006]. The regional mean total mass and aerosol chemical species concentrations over the Bay of Bengal and Arabian Sea during ICARB obtained from HVS measurements and analysis are compared with OPAC estimates of the same in Table 2. TSP from HVS and the mass concentrations of different aerosol species are found to compare very well with the OPAC estimates of the same (Figures 3b and 4). The coefficient of determination (square of the correlation coefficient, R2) between the TSP mass from HVS and Sun photometer AODs at 0.50 μm is found to be 0.8 over the Bay of Bengal and Arabian Sea.
Table 2. Mean Total Aerosol Mass and Species Mass Concentrations Deduced From High-Volume Sampler Measurements in Comparison With Optical Properties of Aerosols and Clouds Model Estimates Over the Bay of Bengal and Arabian Sea During March–May 2006a
Bay of Bengal
Unit is μg m−3. TSP, total suspended particulate mass; HVS, high-volume sampler; OPAC, Optical Properties of Aerosols and Clouds (OPAC).
23.3 ± 12.7
21.0 ± 12.0
25.2 ± 10.5
22.2 ± 10.8
10.2 ± 4.9
9.8 ± 5.0
5.2 ± 2.2
5.0 ± 2.1
1.3 ± 0.9
1.2 ± 0.9
5.0 ± 7.6
4.9 ± 7.5
11.9 ± 6.6
10.9 ± 6.7
10.9 ± 6.0
10.5 ± 5.6
0.4 ± 0.2
0.4 ± 0.2
0.09 ± 0.03
0.09 ± 0.03
 The mass concentrations of water-soluble species, sea salt, black carbon and mineral dust from OPAC and HVS over the Bay of Bengal and Arabian Sea are shown in Figure 4. The trends in the variation of water-soluble, black carbon and mineral dust are similar to that of TSP over the Bay of Bengal. Sea salt concentrations over the Bay of Bengal did not show much day to day variability and the concentrations are only 2–3 μg m−3. Black carbon mass concentrations are also low (<0.5 μg m−3) over the Bay of Bengal. Water-soluble constituents (44%) and mineral dust (48%) dominate and contribute 92% to TSP. Sea salt and black carbon contributed 6 and 2%, respectively, to TSP over the Bay of Bengal. The variation in the mass fractions of aerosol species over Arabian Sea during different days exhibit a distinctly different behavior relative to the variations over the Bay of Bengal. The water-soluble aerosol mass concentration shows similar variation as that of TSP with a mean value of about 5 μg m−3; while sea salt is found to show a large variability ranging from 5 to 30 μg m−3. EC contribution is nearly insignificant and is lower than the Bay of Bengal. Over Arabian Sea, sea salt contributes about 22% to TSP, whereas water-soluble aerosols and mineral dust contribute 30 and 47%, respectively, and black carbon accounts for 1% of TSP.
4.2. Single Scattering Albedo and Asymmetry Parameter
 SSA values can range from 1 (pure scatterer) to 0 (pure absorber). SSA of sulfate and sea salt is 1 at λ = 0.55 μm while the SSA of black carbon is 0.21 at the same wavelength. SSA in the spectral range of 0.25 to 4.0 μm for aerosol models varying from continental clean to maritime polluted [Hess et al., 1998] are shown for 0% and 70% RH in Figure 5. Continental clean aerosol model depicts remote continental locations with or without very low anthropogenic influence [Hess et al., 1998]. Continental average aerosol type corresponds to continental areas influenced by man-made activities. Continental polluted aerosol model represents areas highly polluted by anthropogenic activities. Continental clean model is made up of water-soluble (2600 particles cm−3) and insoluble (0.15 particles cm−3) aerosol components. Continental average and continental polluted models comprise water-soluble, insoluble and black carbon aerosols. Continental average has 7000 water-soluble particles, 0.4 insoluble and 8300 black carbon aerosols per cm3, while continental polluted has 15,700 water-soluble, 0.6 insoluble and 34,300 black carbon particles per cm3. Continental polluted has more than twice the amount of water-soluble aerosols and 4 times higher black carbon aerosol content when compared to continental average aerosol model. Urban aerosol model is for urban areas which have strong pollution. Urban aerosol model has 28,000 water-soluble particles, 1.5 insoluble particles and 130,000 black carbon aerosols per cm3. Maritime clean aerosol model represents undisturbed remote marine regions. Maritime tropical aerosol model has lesser number of water-soluble and sea salt aerosols when compared to maritime clean model. The sea salt number density is lower in maritime tropical model as lower wind speeds are assumed [Hess et al., 1998]. Maritime polluted aerosol model corresponds to a marine region under anthropogenic influence with highly variable amounts of black carbon and water-soluble aerosols [Hess et al., 1998]. Maritime clean model has 1500 water-soluble particles per cm3, 20 sea salt accumulation and 3.2 × 10−3 sea salt coarse particles. Maritime tropical model has 590 water-soluble, 10 sea salt accumulation and 1.3 × 10−3 sea salt coarse particles [Hess et al., 1998]. Maritime polluted model contains 3800 water-soluble particles per cm3 and the same number concentrations of sea salt in accumulation and coarse modes as in maritime tropical, in addition to 5180 black carbon aerosols per cm3.
 When aerosols are dry (0% RH) SSA for maritime tropical and clean aerosol models is high (>0.98) throughout the shortwave spectral range indicating the dominance of scattering aerosols (water-soluble and sea salt). SSA for maritime polluted aerosol model decreases because of black carbon aerosols. SSA for urban aerosol is the lowest among the different aerosols (Figure 5a). SSA decreases almost linearly from continental clean to continental polluted. Continental clean aerosol model is dominated by water-soluble aerosols, while continental average and polluted aerosol models in addition have varying amounts of black carbon which reduces the SSA. In the urban aerosol model black carbon dominates the number mixing ratio (82%) thereby leading to a large reduction in SSA. The sizes of hygroscopic aerosols such as water-soluble and sea salt particles can increase as relative humidity increases; as the radius of scattering aerosols increases the scattering coefficient and SSA will increase as seen at 70% RH (Figure 5b). Asymmetry parameter is found to be higher for maritime aerosol models when compared to continental aerosol models (Figure 5c). The g values will be higher for an aerosol size distribution consisting of bigger particles. The spectral variation of asymmetry parameter for continental models also differs with respect to the maritime models (Figure 5c). Asymmetry parameter also increases when RH increases to 70% (Figure 5d). The measured aerosol properties, however, may have additional components with varying number densities depending on local sources and long-range transport which could modify the aerosol properties when compared to those obtained from aerosol models.
 OPAC estimated single scattering albedo at 0.50 μm is found to be in the 0.91–0.95 range over the Bay of Bengal region with a mean of 0.93 ± 0.01 (Figure 6a). The highest value of SSA (0.95) is observed on 26 March and the lowest value (0.91) is observed on 30 March. SSA estimated using OPAC compares well with the mean value of 0.93 ± 0.03 estimated by Nair et al. . SSA > 0.9 suggests the dominance of scattering aerosols over the Bay of Bengal during March–April 2006. The occurrence of higher SSA is further corroborated by a low (2%) contribution of black carbon to TSP mass. Nair et al.  estimated SSA from the scattering and absorption coefficients measured on board using nephelometer and aethalometer, respectively. SSA was found to vary from 0.96 to 0.84 over the Bay of Bengal [Nair et al., 2008]. SSA was low (0.84) near Port Blair (12°N, 93°E) in central Bay of Bengal, while a high value of 0.98 was obtained over the southern Bay of Bengal (6°N, 89°E) [Nair et al., 2008]. Though the mean SSA values agree well between the present study and Nair et al.'s results, there exist day to day differences. The day to day differences could arise due to the following major reasons: (1) methodology (results in the present study are obtained from the hybrid approach of combining all the chemical composition data and OPAC, while Nair et al. used scattering and absorption coefficients measured by nephelometer and aethalometer, respectively), (2) the present approach includes all the absorbing species (black carbon, mineral dust, and organics) while absorption coefficients reported by Nair et al. pertain only to black carbon, (3) sampling (one aerosol sample of 15–22 h in a day is obtained in our case, while Nair et al. results are obtained from several data points in a day as the measurements are continuous), and finally, (4) uncertainties (the chemical composition which is used in the determination of SSA can be uncertain by about 15% while the uncertainty in SSA derived by Nair et al. can range from 10 to 40% depending on the absolute value of scattering and absorption coefficients). Also, in a small measure the differences in SSA could have occurred due to different wavelengths used as SSA in the current study is at 0.5 μm while Nair et al.  results correspond to 0.55 μm.
 The mean SSA over the Arabian Sea is found to be 0.96 which is higher than the Bay of Bengal SSA (Table 3). The maximum value of SSA is found to be 0.98 on 4 May when the ship was in the northern Arabian Sea (Figure 6a). SSA value over the Arabian Sea is higher than the Bay of Bengal owing to a lower carbonaceous aerosol mass concentration (Table 2). SSA values from previous studies conducted prior to ICARB over the Bay of Bengal and the Arabian Sea regions were found to be lower (Table 3). It should be noted that all the previous studies were conducted over the Bay of Bengal during the northeast winter monsoon (December–March) in contrast to the ICARB cruise which was conducted during the premonsoon (March–April). During the winter monsoon the estimated SSAs are lower because of the dominance of absorbing (carbonaceous) aerosols emanating from fossil fuel and biomass burning mainly from the south Asian region which then get transported across the Bay of Bengal and Arabian Sea [Ramachandran, 2005]. Higher SSA values during the premonsoon season over the Bay of Bengal and Arabian Sea confirm the dominance of scattering aerosols in the size distribution.
Table 3. Mean Aerosol Optical Depth and Single Scattering Albedo at 0.5 μm, Aerosol Radiative Forcing at the Top of the Atmosphere, the Surface, and in the Atmosphere, Along With the Heating Rate Over the Bay of Bengal and Arabian Sea Obtained From the Present Study in Comparison With Earlier Resultsa
Region and Source
Forcing (W m−2)
Heating Rate (K/d)
Aerosol radiative forcing and heating rate are normalized with AOD and given. AOD, aerosol optical depth; SSA, single scattering albedo; TOA, top of the atmosphere; SFC, surface; ATM, atmosphere.
 The day to day variation in asymmetry parameter at 0.5 μm over the Bay of Bengal and Arabian Sea obtained from OPAC model is plotted in Figure 6b. The average value of asymmetry parameter at 0.5 μm is found to be 0.69 ± 0.01 and 0.70 ± 0.01 over the Bay of Bengal and Arabian Sea, respectively. Smaller asymmetry parameters indicate that the aerosol size distribution is dominated by submicron size aerosol particles. 0.5 μm g values at 70% RH are 0.76, 0.77 and 0.75 for maritime clean, tropical and polluted aerosol models, respectively [Hess et al., 1998]; g is <0.7 at 0.5 μm for continental (clean, average and polluted), and urban aerosol models. The g is found to decrease when submicron size particles increase in the aerosol size distribution (Figure 5); maritime polluted has more number of fine mode (water-soluble particles and black carbon) than maritime clean and tropical aerosol models. Similarly continental clean aerosol model has a higher g than the urban aerosol model (Figure 5). Mean SSA at 0.5 μm over the Bay of Bengal and Arabian Sea are found to be 0.93 ± 0.01 and 0.96 ± 0.01, respectively; mean g values at 0.5 μm over the Bay of Bengal and Arabian Sea are 0.69 ± 0.01 and 0.70 ± 0.01. SSA over the Bay of Bengal is found to be lower than maritime aerosol models (0.97–1.00) (Figure 5) indicating the presence of more number of absorbing aerosols than the maritime polluted model. The g is also lower than the maritime models (0.75–0.77) over the Bay of Bengal further supporting the presence of aerosols from long-range transport. SSA over the Arabian Sea is higher than the Bay of Bengal but still lower than the maritime polluted model. SSA over the Arabian Sea on most of the days are in the range of 0.94–0.96 unlike the SSA over the Bay of Bengal (Figure 6). The g over the Arabian Sea is slightly higher than the Bay of Bengal. The differences between the aerosol properties obtained over the Bay of Bengal and Arabian Sea, and maritime aerosol models corroborate the long-range transport of other types of aerosols (Figure 2) in addition to local sources which modified the aerosol optical properties.
Figure 7 presents the spectral dependence of aerosol optical depth, SSA and g estimated using OPAC in the wavelength region of 0.25 to 4.0 μm for select days with high and low AODs over the Bay of Bengal (Figures 7a–7c) and Arabian Sea (Figures 7d–7f). The measured AOD spectra using Sun photometer are also shown along with ±1σ variation from the mean. Figure 7a shows the AOD spectra for 24 March and 3 April when the mean 0.5 μm AODs were 0.89 and 0.19, respectively. Figure 7d shows the AOD spectra obtained over the Arabian Sea on 26 and 22 April when the Sun photometer derived AODs at 0.5 μm were 0.73 and 0.22, respectively. Spectral dependence is found to be more steeper for the days with higher AODs when compared to the days with lower AODs. This is found to be true for all the days when the AODs are higher over both the oceanic regions. No significant variation in SSA and asymmetry parameter is seen for the days with lower and higher AODs, indicating that the size and the chemical composition are more or less similar during the low- and high-AOD conditions, and the higher AODs would have resulted due to an increase in the number density. Both SSA and g are found to decrease as wavelength increases. The enhancement in AOD around 3 μm, and dips in SSA and g (Figures 5 and 7) are attributed to an increase in the absorption characteristics of water-soluble, dust and sea salt aerosols [d'Almeida et al., 1991; Lacis and Mishchenko, 1995], while the optical characteristics of carbonaceous aerosols exhibit a gradual decrease with respect to wavelength. All the maritime aerosol models are seen to exhibit similar features [d'Almeida et al., 1991].
 Ångström exponent α has been used in many studies as a tool to quantify particle size distribution from spectral distribution of AODs and for extrapolating AOD throughout the shortwave spectral region [e.g., Smirnov et al., 2002]. The Ångström power law is rigorously valid for all the wavelengths only if the particle size distribution fits a Junge power law function [Hess et al., 1998]. However, the Junge power law distributions are accurate only over a limited size range, and extrapolation to smaller or larger sizes may introduce significant errors of the ambient aerosol size distribution. The Ångström exponent (α) can be calculated from the spectral distribution of AODs following Ångström power law as
where τ(λ) is the AOD at a particular wavelength λ (in μm) and β is the turbidity coefficient (AOD at 1 μm). Typical values of estimated from AODs measured in the 0.44–0.87 μm wavelength range are found to vary from 1 to 3 for fresh smoke particles, which is dominated by accumulation mode aerosols to nearly zero for the atmosphere dominated by coarse mode aerosols such as dust and sea salt [Eck et al., 2001]. In the atmosphere the aerosol size distribution is rarely unimodal due to varied sources and their formation mechanisms. Therefore, the Ångström relation is not appropriate for all environments and locations [Eck et al., 2001]. When the aerosol size distribution is multimodal, the wavelength dependence of AOD does not follow Ångström power law and shows departure from the linear behavior of ln AOD versus ln λ [e.g., Eck et al., 2001]. The second-order polynomial fit to examine the curvature in the AOD spectra can be written as
where α0, α1, and α2 are constants. Coefficient α2 represents the curvature observed in the spectral distribution of AODs. α2 is <0 when fine mode particles dominate the aerosol size distribution (such as biomass burning, urban or industrial aerosols), and α2 is >0 when coarse mode aerosols (such as dust, sea salt) are dominant or the aerosol size distribution is bimodal with significant relative magnitude of coarse mode particles [Eck et al., 2001]. Though it has been shown that curvatures in AOD spectra can modify the aerosol optical properties, the influence of curvature in spectral AODs on aerosol radiative forcing is unknown.
 Vertical profile of aerosols is another important input required to estimate ARF, as lack of information on the vertical distribution can introduce uncertainty in ARF estimates [IPCC, 2007]. The presence of elevated aerosol layers over high-reflectance surfaces or a scattering layer can enhance the atmospheric forcing [IPCC, 2007]. It is equally important to determine whether the presence of elevated aerosol layers or inclusion of vertical profiles of aerosols can modify the net energy content of the Earth-atmosphere system.
 In the present study sensitivity tests have been conducted to examine (1) the effect of presence of curvatures in AOD spectra on aerosol radiative forcing and heating rate and (2) whether structures in aerosol vertical profiles alter the net atmospheric forcing and heating rate deduced from the top of the atmosphere and surface aerosol radiative forcings and the results are discussed.
4.4. Effect of Curvature in Aerosol Optical Depth Spectra on Radiative Forcing
 Daily mean AOD spectra measured over the Bay of Bengal and the Arabian Sea have been analyzed for curvature effects (Figures 8a and 8b). The α2 is <0 for all days over the Bay of Bengal and Arabian Sea suggesting the dominance of fine mode aerosols, which is in agreement with higher fine mode fraction (>0.60) values obtained [Kedia and Ramachandran, 2008]. Measured AOD spectra in the 0.4–0.875 μm wavelength range are compared with model estimated AOD spectra (section 3.3) and AOD spectra obtained in the 0.25–4.0 μm wavelength region using α2 values in Figure 8 for Bay of Bengal (3 April) and Arabian Sea (23 April). α2 values are found to be −2.02 and −1.91 for 3 and 23 April, respectively. To ascertain the impact of curvatures in AOD spectra on aerosol radiative forcing and heating rate, AODs in the 0.25–4.0 μm range estimated by combining the measurements and model (OPAC) (section 3.3) are replaced with AOD spectra obtained using α2 (curvature effect) while performing radiative forcing calculations (Figure 8). AOD spectra from all the three cases (measured, OPAC estimated and curvature) are found to agree in the 0.4–0.9 μm wavelength range. OPAC estimated AODs are higher at wavelengths less than 0.4 μm, while AODs derived using curvature effect are lower (Figures 8c and 8d). Model estimated AODs beyond 1 μm are found to decrease following Mie theory and wavelength exponent α, while AODs determined using α2 show a steep decline beyond 1 μm. ARFs at TOA, SFC and ATM using aerosol properties from both the cases, namely, model estimated (case 1) and curvature (case 2), do not show significant differences (Figures 8e and 8f). The solar heating rates are found to differ by less than 0.1 K/d, though, the heating rates estimated using curvature effect are lower. Major portion (∼72%) of the solar irradiance reaching the Earth's atmosphere lies in the 0.2–1.0 μm wavelength range [Seinfeld and Pandis, 1998]. Although the AODs are lower beyond 1 μm due to curvature in AOD spectra, as ARF and heating rate depend on both the AODs and the amount of solar radiation (72% of which lies in the 0.2–1.0 μm wavelength range), they do not differ significantly when the shape of the AOD spectra varies. Thus, though the curvature effect can modify the shape of the aerosol distribution and can affect the aerosol optical properties depending on the dominance of either fine or coarse mode aerosols or a mixture of both, the sensitivity study reveals that the curvature effect does not significantly influence the radiative effects of aerosols.
4.5. Influence of Aerosol Vertical Profiles on Radiative Forcing
 Aircraft measurements made during ICARB near the coast and mainland India indicated the presence of elevated aerosol layers [Moorthy et al., 2008; Satheesh et al., 2009]. Aerosol extinction profiles using a micropulse lidar system from Bhubaneshwar (20.2°N, 85.8°E, 25 March 2006) and Chennai (13.1°N, 80.2°E, 3 April 2006) on the east coast, and Trivandrum (8.5°N, 77°E, 23 April 2006) on the west coast were measured during ICARB [Satheesh et al., 2009]. The aerosol extinction coefficients were found to decrease with an 1/e scaling distance of ∼500 km within the marine boundary layer suggesting the rapidly decreasing impact of continental influence; above the marine boundary layer scaling distances were larger and the gradients were shallower [Satheesh et al., 2009]. As simultaneous lidar measured aerosol profiles from the cruise are not available, model aerosol vertical profiles similar to those obtained during the ICARB air segment from micropulse lidar measurements at 0.523 μm [Satheesh et al., 2009] are constructed and utilized. The mean latitude and longitude of the ship positions on 25 March, 3 and 23 April are 16.97° ± 0.09°N, 92.66° ± 0.47°E, 9.99° ± 0.02°N, 82.84° ± 0.45°E, and 11.00° ± 0.01°N, 62.05° ± 0.59°E, respectively. However, for the sensitivity study, we assume that the vertical structure of aerosols over the marine latitude-longitude domain during the above days is similar to those measured on 25 March 2006, 3 April and 23 April from Bhubaneshwar, Chennai and Trivandrum, respectively. The aerosol extinction profiles are then scaled by the respective day's mean AOD (Figure 3a). Up to 65 atmospheric layers from the surface to 100 km with varying resolution can be introduced in SBDART. The vertical resolution of the atmosphere in this study is varied as follows: the resolution is 0.25 km from the surface to 10 km, 1 km from >10 km to 25 km, 5 km between 25 and 50 km, and 10 km between 50 and 100 km.
 Clear sky shortwave aerosol radiative forcing and solar heating rates estimated including aerosol vertical profiles on 25 March, 3 and 23 April are compared with results obtained without including vertical profiles in Figure 9. ARF at the surface is about a factor of 2 higher (−25 W m−2) on 25 March when compared to those obtained on 3 and 23 April. Mean 0.5 μm AODs on 25 March, 3 and 23 April are 0.43, 0.19 and 0.22, respectively (Figure 3a). AOD on 25 March is about 2 times higher than the AODs obtained on 3 and 23 April. These results suggest that the AODs and radiative forcing are nearly linearly related. ARF is found to be higher from above the surface to up to ∼4 km when compared to the forcing obtained without including vertical profile (Figures 9a–9c); above the crossover altitude of 4 km the aerosol radiative forcing obtained including the aerosol vertical profile is lower. The near surface heating rate is about 0.3 K/d on 25 March and decreases to about 0.2 K/d on 3 and 23 April. In the absence of measured vertical profiles the aerosol extinction is distributed on the basis of scale height which is 1 km for the maritime aerosols [Hess et al., 1998]. Heating rate profiles in both the cases (with and without measured vertical profiles) are found to follow the aerosol extinction profiles (Figures 9d–9f). When aerosols are distributed as a function of scale height the heating rate profile is more smoother, whereas the heating rates obtained by including aerosol vertical profiles are found to exhibit structures consistent with aerosol extinction at different altitudes. Heating rates are found to be higher where aerosol layers exist, and exhibit spatial and altitudinal differences. Heating rate at the aerosol layer peak is found to be >0.3 K/d at 3 km on 25 March, 0.2 K/d on 3 April at 4 km while on 23 April the heating rate is <0.15 K/d peaking below 3 km. However, aerosol radiative forcings at TOA, SFC, ATM, and the heating rates obtained with and without aerosol vertical profiles are found to show negligible differences (Figures 9g–9i), thus, ascertaining that the net energy content trapped in the atmosphere remains almost the same with and without vertical profiles, but only its vertical distribution varies. Nevertheless, as simultaneous aerosol vertical profile measurements are not available during the cruise, and as the motivation of the study is to document the day to day variation in net aerosol radiative forcing between the surface and the top of the atmosphere, aerosol radiative forcing at TOA, SFC and ATM are determined and discussed using the simultaneously measured optical, physical and chemical characteristics of aerosols.
4.6. Daily Mean Aerosol Radiative Forcing Over Bay of Bengal and Arabian Sea During Premonsoon
 The daily mean clear sky shortwave aerosol radiative forcings at the top of the atmosphere (TOA), surface (SFC) and atmosphere (ATM) are shown over the Bay of Bengal and Arabian Sea (Figure 10). Aerosol radiative forcings are found to exhibit day to day variability over the Bay of Bengal. These changes are found to be consistent with the variability seen in AODs as the forcings are nearly linearly connected to the AODs. For example, AODs were maximum on 24 March 2006 (Figure 3a) over the Bay of Bengal region. On 24 March the cruise was at 17°N, 87°E in the middle and interior region of the Bay of Bengal. The higher AODs were attributed to the transport of pollutants from the continent [Kedia and Ramachandran, 2008]. TOA forcing on 24 March is found to be the highest at −28 W m−2; SFC forcing is about −52 W m−2 resulting in an atmospheric warming of 24 W m−2. As SSA values are higher than 0.90 over the Bay of Bengal both the top of the atmosphere and surface forcings increase (more negative), thus, resulting in lesser magnitude of atmospheric warming. In contrast, strongly absorbing aerosols can absorb the radiation reflected upward from the lower layers of the atmosphere and the ocean surface, make the TOA forcing less negative or even positive [e.g., Podgorny and Ramanathan, 2001; Ramachandran, 2005], increase the SFC forcing, thereby leading to a higher (more positive) atmospheric warming. In the present study, the average aerosol radiative forcing over the Bay of Bengal at the TOA and at the surface are found to be −12.0 ± 5.4 and −22.4 ± 9.8 W m−2, respectively. The average atmospheric warming is 10.4 ± 4.6 W m−2 over the Bay of Bengal during March–April 2006.
 Over the Arabian sea the forcing is found to be highest (−30.0 and −45.8 W m−2 at the TOA and SFC, respectively) on 26 April when the ship was moving near the western coast of India. AODs were highest on 26 April over the Arabian Sea (Figure 3a). The forcing decreased very sharply as the ship moved into the central part and the minimum value of forcing (−1.2 and −2.0 W m−2 at TOA and SFC, respectively) is observed on 1 May in the central Arabian Sea when the AODs were low. The average radiative forcings over the entire Arabian Sea are found to be −10.5 ± 6.8, −15.8 ± 10.7, and 5.3 ± 3.9 W m−2 at the TOA, SFC and ATM, respectively. The aerosol radiative forcings at all the three altitude levels over the Arabian Sea are found to be lower than that of the Bay of Bengal. The lower radiative forcings are expected because of lower AODs and higher SSAs over the Arabian Sea when compared to the Bay of Bengal.
4.7. Climate Implications and Comparison
 Daily heating rates over the Bay of Bengal and Arabian Sea during the premonsoon season of 2006 are plotted in Figure 10c. The heating rates are >0.1 K/d over the Bay of Bengal, while over the Arabian Sea heating rates are lower than 0.1 K/d on a few days concurrent with lower aerosol content (Figure 3a). Heating rate is maximum (0.67 K/d) on 24 March over the Bay of Bengal when AODs are also highest, while heating rate is the lowest at 0.14 K/d on 4 April 2006 (Figure 10c). Highest heating rate of 0.44 K/d over the Arabian Sea occurred on 26 April. The mean solar heating rate over the Bay of Bengal during ICARB is 0.3 K/d, which is twice larger when compared to Arabian Sea (shown as horizontal lines in Figure 10c). A comparison of aerosol radiative forcing and heating rate from the present study with earlier results over the Bay of Bengal and Arabian Sea can bring out the seasonal variations in aerosol radiative impact over these oceanic regions. The heating rates determined in the current study and in all the earlier studies (Table 3) correspond to a ΔP of 300 hPa. Assuming a different thickness (ΔP) would not change the conclusions, but the absolute magnitudes of the heating rates would only vary; that is, a higher ΔP would reduce the heating rate value. As earlier studies were conducted during different years and different seasons, for comparison, aerosol radiative forcing and heating rates obtained from earlier studies and ICARB have been normalized with AODs (Table 3). The comparison clearly shows that the heating rates obtained over the Bay of Bengal and Arabian Sea during ICARB are lower than the heating rates obtained earlier (Table 3). The mean atmospheric warming over the Bay of Bengal and the Arabian Sea in the present study is the lowest ever obtained in the last decade (1996–2006) (Table 3) suggesting the dominance of scattering aerosol species during premonsoon. The earlier studies conducted during the winter monsoon season when the winds are from the polluted northern hemisphere were marked by the occurrence of higher AODs and lower SSAs, which gave rise to higher ATM (more positive) aerosol radiative forcings and heating rates. The seasonal and spatial variations in aerosol radiative forcing and heating rate over the oceanic regions surrounding India will be useful in the radiative and climate impact assessments.
 Spatial and temporal heterogeneity in aerosol radiative forcing and heating rate are investigated over the Bay of Bengal and Arabian Sea for the premonsoon season of 2006. Aerosol optical depths, total aerosol mass and chemical composition in combination with a radiative transfer model are used to estimate the radiative effects of aerosols. The major findings of the study are as follows.
 1. Aerosol optical depths over the Bay of Bengal are higher than the Arabian Sea during the premonsoon season of 2006, and are found to exhibit day to day variations. The total suspended particulate (TSP) mass concentrations are in the range of 5–47 μg m−3 over the Bay of Bengal and Arabian Sea. Over the Bay of Bengal water-soluble and mineral dust aerosols contributed more than 90% to TSP, while sea salt and BC (or EC) contributed less than 10% to TSP. In contrast over the Arabian Sea water-soluble and mineral dust contributed 77% while sea salt contribution was higher than the Bay of Bengal at 22%. EC contributed 1% to TSP mass over the Arabian Sea.
 2. No significant variation in single scattering albedo and asymmetry parameter is seen for the days with lower and higher aerosol optical depths over the Bay of Bengal and Arabian Sea during ICARB, suggesting that, the size distribution and chemical composition of aerosols are similar during low- and high-AOD conditions and, the higher AODs would have resulted due to an increase in the number density.
 3. A sensitivity test revealed that the aerosol radiative forcing and heating rate do not differ significantly when curvatures in aerosol optical depth spectra exist, because the radiative effects of aerosols depend both on the AODs and the incoming solar radiation, 72% of which lies in the 0.2–1.0 μm wavelength range.
 4. Another sensitivity study showed that the presence or the absence of aerosol vertical profiles does not significantly modify the net energy content of the Earth-atmosphere system at the top of the atmosphere, at the surface and in the atmosphere.
 5. The forcing is maximum over the Bay of Bengal on 24 March with the largest atmospheric forcing of about 24 W m−2, while the minimum atmospheric absorption is found to be 4.8 W m−2 on 4 April. The average aerosol radiative forcings over the Bay of Bengal are estimated to be −12.0, −22.4 and 10.4 W m−2 at TOA, SFC and ATM.
 6. Over the Arabian Sea the forcing is found to be maximum when the ship was moving near the southern Indian peninsula. The average aerosol radiative forcings are found to be −10.5, −15.8, and 5.3 W m−2 at TOA, SFC and ATM over the Arabian Sea.
 7. The solar heating rates are found to be higher than 0.1 K/d over the Bay of Bengal. Lower heating rates are obtained over the Arabian Sea where AODs are lower. Average heating rate over the Bay of Bengal (∼0.3 K/d) is twice higher than the Arabian Sea.
 8. A comparison of aerosol radiative forcing and heating rate with previous results obtained over the Bay of Bengal and Arabian Sea revealed that the atmospheric warming and the associated heating rate are lower during March–May 2006. Most of the earlier studies were conducted during the winter season when pollutants are transported from the northern hemisphere. Higher AODs accompanied with lower SSAs gave rise to higher ATM warming (more positive) and higher heating rate in winter, while during premonsoon TOA and SFC forcings are higher (more negative) due to the dominance of scattering aerosols resulting in a lower ATM warming (less positive) and lower heating rate. Thus, the aerosol radiative forcing results obtained from ICARB will be useful in assessing the regional and seasonal climate impacts over the Bay of Bengal and Arabian Sea.
 We thank K.K. Moorthy, Project Director, ICARB, C.B.S. Dutt, ISRO-GBP Program Office, ISRO Headquarters, Bengaluru for the efficient planning and conduct of the campaign, and for the support. We are grateful to the National Centre for Antarctic and Ocean Research (NCAOR) and the Department of Ocean Development for giving us an opportunity to sail and conduct measurements on board Sagar Kanya. The air back trajectories are obtained using HYSPLIT (version 4) model from http://www.arl.noaa.gov/ready/hysplit4.html. We thank Rohit Srivastava for his help in performing radiative forcing calculations using aerosol vertical profiles.