Anthropogenic sulphate aerosol from India: estimates of burden and direct radiative forcing

Anthropogenic sulphate aerosol from India: estimates of burden and direct radiative forcing

Atmospheric Environment 33 (1999) 3225}3235 Short Communication Anthropogenic sulphate aerosol from India: estimates of burden and direct radiative ...

211KB Sizes 2 Downloads 28 Views

Atmospheric Environment 33 (1999) 3225}3235

Short Communication

Anthropogenic sulphate aerosol from India: estimates of burden and direct radiative forcing Chandra Venkataraman *, Bharadwaj Chandramouli , Anand Patwardhan Centre for Environmental Science and Engineering, Indian Institute of Technology, Bombay, Powai, Mumbai 400 076, India School of Management, Indian Institute of Technology, Bombay, Powai, Mumbai 400 076, India Received 27 August 1997; accepted 18 March 1998

Abstract A one-box chemical-meteorological model had been formulated to make preliminary estimates of sulphate aerosol formation and direct radiative forcing over India. Anthropogenic SO emissions from India, from industrial fuel use and  biomass burning, were estimated at 2.0 Tg S yr\ for 1990 in the range of previous estimates of 1.54 and 2.55 Tg S yr \ for 1987. Meteorological parameters for 1990 from 18 Indian Meteorological Department stations were used to estimate spatial average sulphate burdens through formation from SO reactions in gas and aqueous phase and removal by dry  and wet deposition. The hydrogen peroxide reaction was found dominating for undepleted oxidant-rich conditions. Monthly mean sulphate burdens ranged from 2}10 mg m\ with a seasonal variation of winter}spring highs and summer lows in agreement with previous GCM studies. The sulphate burdens are dominated by sulphate removal rates by wet deposition, which are high in the monsoon period from June}November. Monthly mean direct radiative forcing from sulphate aerosols is high (!3.5 and!2.3 W m\) in December and January, is moderate (!1.3 to!1.5 W m\) during February to April and November and low (!0.4 to!0.6 W m\) during May to October also in general agreement with previous GCM estimates. This model, in reasonable agreement with detailed GCM results, gives us a simple tool to make preliminary estimates of sulphate burdens and direct radiative forcing.  1999 Elsevier Science Ltd. All rights reserved. Keywords: Global climate; Sulphate aerosol; India; Anthropogenic SO emissions; Chemical-meteorological box model 

1. Introduction The signi"cance of climate forcing by anthropogenic sulphate aerosols has recently been recognised, along with its strong regional variation (Charlson et al., 1991, 1992; Lelieveld and Heintzenberg, 1992; Kiehl and Briegleb, 1993). The sensitivity of direct climate forcing by aerosols is strongly dependent on the sulphate aerosol mass because its hygroscopicity promotes water ac-

* Corresponding author.

cretion, and increased scattering, at relative humidities of interest in the troposphere (Pilinis et al., 1995). Anthropogenic sulphate aerosol is formed primarily from reactions of sulphur dioxide in the atmosphere with oxidants like hydroxyl radicals (gas phase), ozone, hydrogen peroxide and oxygen (aqueous phase in clouds), with the aqueous phase reactions contributing about two-thirds of the yield in model simulations (Feichter et al., 1996). The conversion of sulphur dioxide to sulphate aerosol is dependent on various factors including oxidant concentrations, cloud pHs, dry/wet deposition #uxes of SO  and meteorological parameters including temperature,

1352-2310/99/$ - see front matter  1999 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 2 - 2 3 1 0 ( 9 8 ) 0 0 1 4 0 - X


C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

relative humidity and cloud cover (Langner and Rodhe, 1991; Feichter et al., 1996). The column burden of this aerosol, and the consequent radiative forcing, is dependent on its wet and dry deposition rates. Estimates of global sulphate distributions from three-dimensional global scale chemical/transport models (Langner and Rodhe, 1991; Pham et al., 1995; 1996; Chin et al., 1996 a,b; Feichter et al., 1996) provide a global distribution of sulphate aerosol. While these models provide extremely detailed results, their disadvantage is the requirement of high performance computation and of very intensive databases of emission inventories and meteorological/chemical parameters, which are not often available for recent years and which may not re#ect emission factors suitable to a given country. In order to re"ne regional estimates of sulphate burdens and direct radiative forcing, we have used new estimates of aggregate emissions for India along with meteorological parameters from 18 stations in a simple box model incorporating chemistry and removal mechanisms for SO and sulphate. We examine the relative  importance of di!erent SO oxidation pathways and  estimate monthly mean sulphate burdens, optical depths and direct radiative forcing over the Indian subcontinent. These are compared with GCM estimates for the region and available measurements for India.

2. Model description A "rst estimate of sulphate aerosol yield and contribution to radiative forcing is made using regional emissions and atmospheric parameters for the Indian subcontinent. An SO emission inventory is compiled from aggregate  industrial fuel consumption and biomass burning along with appropriate emission factors. A box model is used to estimate sulphate aerosol formation from gas and aqueous phase reactions of SO and calculate an average  &&yield''. Removal mechanisms include advection, dry deposition of SO and wet deposition of sulphate. The  direct radiative forcing is calculated from a sulphate aerosol burden and a mass scattering e$ciency corrected for relative humidity variations. The spatial average sulphate yield for India is estimated from point yields using meteorological data from 18 stations spread uniformly over India averaged by a Fourier amplitude sensitivity test which varies the input chemical and meteorological parameters over their monthly ranges. 2.1. Emissions An emission inventory was compiled combining fuel use estimates for 1990 with appropriate emission factors. 1990 industrial coal and petroleum use (Mitra, 1992) were compiled (Table 1) along with average sulphur contents of 0.6% for Indian hard coals and 1.3% for

Table 1 Anthropogenic SO emissions from India for 1990  Source

Coal Steel Power Railways Cement Sp. Iron Fertiliser Coke Others Lignite Subtotal Petroleum LPG Motor Gas. Kerosene Aviation gas H S Diesel Fuel oils Subtotal Biomass Wood Cattle Dung Agri. Residue Forest Biomass Grassland Subtotal Total

Consumption, MT yr\

SO emitted,  Tg S yr\

22.3 135.7 6.5 11.5 1.0 5.5 3.5 31.7 11.0 228.7

0.131 0.794 0.038 0.067 0.006 0.032 0.006 0.185 0.107 1.366

2.3 3.5 8.2 1.8 20.7 8.8 45.3

0.000 0.001 0.001 0.000 0.005 0.159 0.166

252.1 106.9 99.2 26.7 5.3 490.1

0.284 0.083 0.077 0.021 0.004 0.470 2.0

Energy use from: Mitra (1992), Kaul (1993), Kaul and Shah (1993), Sinha and Joshi (1997). Emission factors from: Spiro et al. (1992), Andreae et al. (1989), Mazumdar et al. (1985).

lignite (Mazumdar et al., 1985) and sulphur emission factors of 0}1.8% S (Spiro et al., 1992) for petroleum products to obtain annual average SO emissions of 1.37  Tg S yr\ (68%) from coal and 0.17 Tg S yr\ (9%) from petroleum products. 1989}90 estimates of biomass-burning (Kaul, 1993; Kaul and Shah, 1993; Sinha and Joshi, 1995) along with an emission factor of 0.32;10\ mol Smol\ C burnt (Andreae et al., 1988), and an assumed biomass carbon content of 45% (Mitra, 1992), resulted in 0.47 Tg S yr\ (23%) of SO emissions. Total SO   emissions from industry and biomass burning are 2.0 Tg S yr\ for India for 1990. Spiro et al. (1992) gave a 1980 estimate of 2.8 Tg S yr\ anthropogenic and natural SO  emissions from India, of which 76.1% was apportioned to industrial activities and 2.2% to biomass burning, giving a "gure of 2.2 Tg S yr\ for 1980 from industry and biomass burning. Accounting for the approximate factor of 2 increase in industrial fuel use in India between

C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

1980 and 1990, the Spiro et al. (1992) estimate would scale to 4.4 Tg S yr\ for 1990, in comparison to our estimate of 2.0 Tg S yr\. The large di!erence in the two estimates comes from the di!erent assumptions of hardcoal sulphur contents of 2.3% (Spiro et al., 1992) and 0.6% in the present work (Mazumdar et al., 1985) which we believe is a more realistic "gure for Indian coals. More recent estimates of sulphur emissions from India include 1.54 Tg S yr\ for 1987 (Kato and Akimoto, 1992; Akimoto and Narita, 1994) (25% lower than our estimate) using a 0.63% S content for coal and accounting only for wood and baggase forms of biomass burning. The estimate of 2.55 Tg S yr\ for 1987}1988 (Arndt et al., 1997) (20% higher than our estimate) compares most closely with our source accounting and includes all biomass sources (including dried animal dung) with the discrepancy arising from their assumption of 1% S content of coal (as opposed to 0.6% in the present work) and their inclusion of emissions from shipping. 2.2. Box model for sulphate yield The formulated box model covers an area (L;W) of 7;10 m (8}36 N and 70}92 E), including the landmass of India and surrounding oceans, with an average mixed layer height of 0.5}2 km (Charlson et al., 1992) and with uniformly mixed constituents. Sinks of SO include  reaction to form sulphate, advection and dry deposition. Wet deposition of SO is neglected as earlier studies  (Feichter et al., 1996; Langner and Rodhe, 1991) indicate that it is a minor sink mechanism accounting for less than 10% of SO deposition. As the time scale of SO aqueous   oxidation is almost instantaneous, particularly with H O availability, the SO aqueous oxidation lifetime is    limited by the time of delivery of an SO containing air  parcel to a cloud (Langner and Rodhe, 1991). A SO  balance is given by


1 v Q"

vcosh vsinh # ¸ =


where Q is the SO emission rate (Tg S s\), < the  volume of the free troposphere over A (m), C the steadystate SO concentration (Tg S m\), K the "rst-order  -& rate constant for SO oxidation by hydroxyl radicals  (s\), s the time scale of delivery of a SO containing air   parcel to clouds (s), v the SO dry deposition velocity (m   s\), h the mixed volume height (m), v the average advection velocity (m s\) and h the average wind direction. s is the same quantity de"ned as s by Langner and   Rodhe (1991) and given as q "(1!b)q  



with s being the time scale between successive cloud  encounters, the monthly mean values of which range 12}56 hr for 0}403N (Lelieveld et al., 1989). The fraction of the emitted SO reaching clouds, F, is  given as the ratio of the delivery rate to cloud and the emission rate 1/q  F" . K #1/q #v /h#(v cos h/¸#v sin h/=) -&  


This fraction reacts by aqueous phase reactions which occur in clouds. The sulphate yield is given as the product of F and the fractional SO in cloud conversion:  >"F .

bHSO [KO #KO #KH O ] 


1 bHSO [KO #KO #KH O ]#      q R¹ 


where KO , KO , KH O (s\) are the pseudo "rst-order     rate constants for aqueous oxidation by oxygen, ozone and hydrogen peroxide, s (s) is the in-cloud residence  time, and HSO (mol l\ atm\) is the Henry's law  constant for SO dissolution. The aqueous reaction rate  depends upon oxidant availability and tends to be very high with undepleted H O availability.   2.3. Chemistry Reactions of sulphur dioxide to sulphate include gasphase oxidation by hydroxyl radicals (Meng and Seinfeld, 1994) and aqueous phase oxidation by oxygen (catalysed by Fe> and Mn> ), ozone and hydrogen peroxide (Martin and Damschen, 1981; Maahs, 1983; Ibusuki and Takeuchi, 1987). The reactions are assumed to occur in oxidant rich conditions with no oxidant depletion. SO and aqueous phase oxidants are assumed  to dissolve instantaneously under equilibrium conditions governed by temperature dependent Henry's Law coe$cients. Under these conditions, the contribution of each oxidant to sulphate yield is directly proportional to the oxidant concentration and the cloud volume fraction, b. We used an ozone concentration range of 0}100 ppb which covers the range of surface ozone concentrations measured at various locations in India (Khemani et al., 1995) and would somewhat overestimate average ozone concentrations in the mixed layer. Cloud pH is often approximated by precipitation pH which ranges 5}7 in India (Parashar et al., 1996). Table 2 lists average oxidant and catalyst concentrations used in the calculation. The rate equations and reaction rate constants are listed in Appendix A. 2.4. Removal processes SO is removed by dry deposition and advection.  Average surface SO dry deposition velocities of 0.6 and 


C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

Table 2 Values of chemical and meteorological parameters used for sulphate yield Parameters

Cloud volume fraction, b (%) Mixing height, h (m) Wind velocity, < Temperature, ¹ OH concentration (molecules/cm) Fe> concentration (M)

Range of values



IMD (1990)

500}2000 m Charlson et al. (1992)

Table 3 Constants used in the radiative forcing estimate Symbols


F  ¹ 

Top of atmospheric radiative #ux 1370 W m Fraction of light transimitted by 0.76 above aerosol layer Fractional albedo of underlying 0.15 surface Backscatter fraction 0.29

R  b

Met. Data Met. Data 3-10;10

IMD (1990) IMD (1990) Meng and Seinfeld


Mn> concentration (M)


O concentration  (ppb in air) H O concentration   (ppt in air) Cloud pH


Ibusuki and Takeuchi (1987) Ibusuki and Takeuchi (1987) Khemani et al. (1995)


Feichter et al. (1996)


Parashar et al. (1996)

0.8 cm s\ for deposition over land and ocean (Langner and Rodhe, 1991, Feichter et al., 1996) and a global average value of 0.5 cm s\ (Charlson et al., 1992) have been used in previous studies. As over 80% of the box area is covered by land, we use a range of 0.5}0.7 cm s\ (average of 0.6 cm s\) in this work. Average advection velocities, obtained from surface wind velocity measurements made at meteorological stations by the Indian Meteorological Department (IMD), range from 1.3}2.7 m s\. As these are likely to underestimate longrange transport, we use average advection velocities in the free-troposphere over India which range from 0}15 m s\ for June}August and from 5}35 m s\ for December}February (Hastenrath, 1985). Average free-tropospheric wind speeds over India for January and July from a global circulation model (GCM) also corroborated these ranges (Boucher, 1997). 2.5. Radiative forcing Direct radiative forcing is estimated assuming an optically thin layer of aerosol where multiple re#ections within the aerosol layer are neglected (Charlson et al., 1991, 1992). The change in radiative forcing in a column of air integrated from the surface of the earth to the top of the atmosphere, as a result of the change in planetary albedo occurring in the cloud free area of the atmosphere, is given by (Charlson et al., 1991, 1992) DF"!1/4 F (1!A )DR 2 



From: Charlson et al. (1991, 1992).

where *F is the change in forcing (W m\), ( )F the  2 incident top of the atmosphere radiative #ux, A the  fractional areal cloud cover, and *R the change in planetary albedo. This assumes that planetary albedo is enhanced only in cloud-free regions, while it has recently been shown (Boucher and Anderson, 1995) that direct radiative forcing in cloudy regions may be 25% as e!ective as in cloud-free regions. The assumption of an optically thin aerosol allows *R to be related to the aerosol optical depth as DR "2¹(1!R )bd (6)  where R is the mean albedo of the underlying surface,  d the sulphate aerosol optical depth, ¹, the fraction of light transmitted by the atmospheric layer above the aerosol layer and b, the average upward scattered fraction. Typical values for these optical parameters (Charlson et al., 1992) are summarised in Table 3. The sulphate aerosol optical depth is a product of its mass scattering e$ciency a (m g\) and its areal column burden, B (g m\): d"aB


with the areal column burden given as (Charlson et al., 1991, 1992): QSO >¸SO   B" A


where, QSO is the sulphur dioxide emission rate (con verted to sulphate), > the fractional sulphate yield, ¸SO the sulphate atmospheric lifetime and A, the area of  the box. ¸SO depends upon the sulphate dry and wet  deposition rates with wet deposition, by rainout and washout, being the primary sink. The mass scattering e$ciency of dry sulphate aerosol (relative humidity (50%) has a mean value of 5 m (g SO )\, which  increases by a factor of 1.7 with water accretion at higher relative humidites of about 80% (Charlson et al., 1991; 1992). Pilinis et al. (1996) showed a large increase in the mass scattering e$ciency of a bimodal aerosol of global

C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

average composition with increase in its sulphate mass. As we estimate only the sulphate forcing in this work, we use the suggested empirical enhancement factor of 1.7 for RH increase from 50 to 80% to calculate the sulphate mass scattering e$ciency. 2.6. Sulphate yield averaging The Fourier amplitude sensitivity test (FAST), used for global sensitivity analysis of non-linear models (Cukier et al., 1973; McRae et al., 1982), is used to obtain average sulphate yields from ranges of input parameters to the yield equation. FAST associates each uncertain parameter with a speci"c frequency in the Fourier transform space of the system. The system sensitivities are then determined by solving the system equations for discrete values of the Fourier transform variable and computing the Fourier coe$cient associated with each parameter. Computer codes are available for applying FAST on user-de"ned subroutines (McRae et al., 1982). The parameters are normally distributed about the mean with an additive search curve for small range or uncertainty and an exponential search curve for a range exceeding an order of magnitude. Ten parameters were included in the yield averaging scheme including global average ranges for dry deposition velocity, oxidant concentrations and cloud pH (Table 2) and monthly ranges of meteorological parameters for India, discussed in the following section.

3. Results and discussion. Monthly mean parameters for India 3.1. Meteorological parameters Monthly mean meteorological parameters, including temperature, surface wind velocity, relative humidity (RH), rainfall and areal cloud cover, were obtained for the year 1990 from the Indian Meteorological Department (IMD), Pune, India, from 18 meteorological stations which uniformly covered the subcontinent (Fig. 1). These data are measured according to WMO guidelines and all parameters (except rainfall) are measured twice a day at 03 : 00 GMT (08 : 30 IST) and 12 : 00 GMT (17 : 30 IST). Temperature is measured using dry bulb, wet bulb and maximum}minimum thermometers installed in a Stevenson screen. RH is derived from the dry and wet bulb temperatures from a hygrometric chart. Wind speed is measured by a manual or an automatic anemometer and wind direction using a standard wind vane. Cloud cover is estimated visually, for eight sections of the sky, as the extent of low, medium and high cloud in each section in 0.1 Oktas, with 1 Okta representing oneeighths of the sky. Rainfall is measured continuously


using a standard recording rain guage (SRRG). Frequency distributions were plotted of the daily mean parameter values from the 20 meteorological stations, and the monthly mean and standard deviation of each parameter calculated for the whole Indian subcontinent. Monthly mean meteorological parameters for India for 1990 (Figs 2}5) show surface average temperatures between 20.6 and 30.83C and RH between 50}80%. There are large monthly variations in areal cloud cover (23}84%) and rainfall (0.02}27.9 cm) with an anomalously high rainfall of 18.7 cm in May. 3.2. Cloud volume fractions Cloud volume fractions were estimated from the monthly mean areal cloud cover data using average cloud heights for six cloud types (Lelieveld et al., 1989) at the latitudes of interest (8}403N). As the available areal cloud covers for India do not distinguish between cloud types, equal occurrence of all six cloud types was assumed to calculate average cloud heights and clouds were assumed to occur throughout the troposphere (12 km) to calculate monthly fractional cloud volumes. The calculated fractional cloud volumes (Fig. 6) range from 8}13% between November and April (low of 8% in January) and compare well with northern hemisphere averages estimated by Lelieveld et al. (1989). From May to September our cloud volume estimates (18}30% with a high of 30% in July) are higher by factors of 1.5}2.5 than the NH averages, an expected result possibly from regional monsoon activity. Our regional averages estimated here span a wider range as expected than global averages of Lelieveld et al. (1989) for 0}403N. Future cloud volume estimates must be re"ned to account for di!erences in seasonal occurrence of di!erent cloud types. 3.3. Sulphate aerosol lifetimes Sulphate aerosol is removed from the atmosphere by wet and dry deposition with the former being the predominant mechanism. A sulphate dry deposition velocity of 0.2 cm s\ (Langner and Rodhe, 1991; Feichter et al., 1996) is used which would result in a removal rate of 10\ s\ and a corresponding lifetime of 11.57 day (mixed layer height of 0.5}2 km) in the absence of wet deposition. Regional sulphate wet deposition #uxes can be estimated for India using the IMD rainfall data (Fig. 3) and appropriate scavenging ratios. Savoie et al. (1987) reported a correlation between NSS sulphate scavenging ratio and event rainfall for Miami from multiple regression analysis. This correlation was used along with monthly mean rainfalls, to generate monthly mean scavenging ratios which ranged 255 in July (maximum rainfall) to 760 in January (minimum rainfall) with an annual average of 471. Monthly mean SRs for NSS sulphate


C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

Fig. 1. Indian Meteorological Department stations, located uniformly over the subcontinent, from which data were obtained for 1990.

varied by upto a factor 7 for Ireland and Miami (Savoie et al., 1987; Galloway et al., 1993) over the period of an year with an average of 290 for Miami and 527 for Ireland. The sulphate life time is calculated as: 1 1 1 ¸SO " " #  K K K  


where, k and k are the respective dry and wet depos  ition rate constants (s\). Monthly mean sulphate lifetimes (Fig. 7) range from 1.6 day in July (very e$cient wet deposition with the maximum rainfall of 279 mm) to 8.2

day in January (minimum rainfall of 2 mm) and compare within a factor of 2 with global average values (3.2}6.1 day) obtained from a tropospheric sulphate cycle simulation study assuming dry deposition and in-cloud scavenging as the predominant removal mechanisms (Langner and Rodhe, 1991). Previous annual average sulphate lifetime estimates of 4.4 day (Feichter et al., 1996) and 4.6}5.2 day (Langner and Rodhe, 1991) for the northern hemisphere lie within that estimated in the present work (3.6$2.3 day). Extensive measurements are needed to develop region speci"c SRs for India to re"ne these estimates.

C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

Fig. 2. Monthly mean temperatures for India for 1990, from Indian Meteorological Department stations.


Fig. 5. Areal cloud covers for India for 1990 exhibit a strong seasonal variation with highs of 60}84% in the monsoon months of May to October.

Fig. 3. Monthly mean rainfall for India for 1990, from Indian Meteorological Department stations. An anomalously high rainfall of 18.7 cm was recorded in May. Fig. 6. Monthly mean cloud volume fractions calculated from IMD areal cloud covers and average cloud heights from Lelieveld et al. (1989). The di!erences between our regional estimates and global average estimates for 0}403N, from May to September, arise possibly from regional monsoon activity over India.

Fig. 4. Monthly mean relative humidity (RH) for India for 1990, from Indian Meteorological Department stations. RHs are within the range of 50}80% observed in the global troposphere.

3.4. Monthly mean sulphate yields and burdens The monthly mean sulphate yield (Fig. 8), averaged spatially using parameters from the 18 meteorological

stations, varied from a low of 0.50 in March and April to a high of 0.72 in July with an annual average of 0.61$0.13. It is controlled by the fraction of SO de livered to cloud which depends on the available cloud volume fraction and the time between successive cloud encounters (Eq. (2)). As we assume undepleted oxidant availability, the H O reaction predominates and com  pletely converts available SO in a fraction of the cloud  residence time. The seasonal trends in the yield are directly related to the cloud volume fraction available for the aqueous phase reactions which is highest in the monsoon season. The integrated column sulphate burdens (Fig. 9) follow those in the sulphate lifetimes and are lowest in May}November (2}3 mg m\) when high rainfalls lead to e$cient sulphate wet scavenging. During the low rainfall period (December}April) sulphate burdens


C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

Fig. 7. Monthly mean sulphate lifetimes from dry and wet deposition of sulphate aerosol show lows from May to November (very e$cient scavenging by monsoon rains) and highs from December to April (dry period). Shown for comparison are global average lifetimes from a sulfur cycle simulation (Langner and Rodhe, 1991) which show more moderate variation.

Fig. 9. Monthly mean sulphate burdens show the same seasonal trend as the lifetimes with an annual variation that is controlled by sulphate wet deposition.

Fig. 8. Monthly mean sulphate yields show highs from May to September when large cloud volume fractions favour sulphate formation by aqueous reactions. The assumption of oxidant rich conditions and undepleted hydrogen peroxide availability leads to high conversion and the yield is limited by the cloud availability rather than the reaction rates.

Fig. 10. Sulphate optical depths follow trends in the sulphate lifetimes and burdens with lows during May}November and highs from December to April.

range from 4}10 mg m\. Annual average sulphate burden is 4.0$2.8 mg m\ in agreement with previous model results for this region (Feichter et al., 1996; Pham et al., 1996). The seasonal trend in sulphate burdens agrees with some recent GCM results (Chin et al., 1996a; Pham et al., 1995) who report, for India, sulphate concentrations of 250}1000 pptv in January higher than those of 50}500 pptv in July. Introducing a scheme for seasonal oxidant availability would further re"ne these yield and burden estimates. 3.5. Sulphate optical depths and direct radiative forcing Sulphate optical depths are calculated as the product of the burdens and the RH corrected sulphate mass

scattering e$ciency (Eq. (7)). The sulphate optical depths (Fig. 10) tend to follow the sulphate lifetimes and burdens with a seasonal variation of lows during May}November (0.01}0.02) and highs from December to April (0.03}0.06) in reasonable agreement with previous model estimates of 0.01}0.03 for India in July as summarized by Kiehl and Rodhe (1995). As we expect that monthly anthropogenic SO emissions would be invariant, monthly variations in  sulphate formation and sulphate removal by wet deposition explain the trends in the sulphate optical depth seen here. Comprehensive measurements of sulphate aerosol size-distributions for India are yet to be made from which sulphate optical depths can be estimated for comparison. The existing measurements for India are of total aerosol optical depths of 0.1}0.8 at "ve stations in

C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

Fig. 11. Monthly mean sulphate direct radiative forcing also shows a seasonal variation with highs in December and January, moderate values during February to April and November and lows during May to October.

India (Krishna Moorthy et al., 1993) and of 0.18}0.6 near the coast (Jayaraman et al., 1997) which have a prominent peak in March}April and a minimum during December}February (Krishna Moorthy et al., 1993). The summer maximum has been attributed to the increased mechanical production of natural aerosols and the winter low to aerosol removal from NE-monsoon rains, particularly for inland midlatitude stations. The present model would need to include all other aerosol sources (black carbon, natural dust aerosols, organic aerosols) before a meaningful comparison can be made with measurements of total aerosol optical depth and its trends. Monthly mean direct radiative forcing from sulphate aerosols is high (!3.5 and !2.3 W m\) in December and January, is moderate (!1.3 to!1.5 W m\) during February to April and November and low (!0.4 to!0.6 W m\) during May to October (Fig. 11). Once again, the forcings follow the seasonal trends in sulphate lifetimes and burdens with an annual average forcing (the average of the 12 monthly means) of !1.1$1.0 W m\. Coupled chemical/radiation models for sulphate aerosols have suggested global mean values of 0.3}1.2 W m\ with local values upto !10 W m\ over industrial regions (Kiehl and Rodhe, 1995). Our results are in general agreement with previous estimates which include a !1.1 W m\ NH average (Charlson et al., 1992), a !0.3 W m\ global average (Kiehl and Briegleb, 1993) and a 0 to !0.5 W m\ forcing over India in July (Kiehl and Rodhe, 1995).

4. Conclusions A single-box chemical}meteorological model was used to make a preliminary estimate of sulphate aerosol


formation and radiative forcing from regional anthropogenic SO emissions from India. Annual average sulphur  dioxide emissions for India from industrial fuel and biomass burning were estimated at 2.0 Tg S yr\ for 1989}1990. Monthly mean sulphate burdens ranged from 2}10 mg m\, with a seasonal variation of winter-spring highs and summer lows controlled by sulphate removal by wet deposition, in reasonable agreement with previous model estimates for this region (Feichter et al., 1996; Pham et al., 1996). The annual average direct radiative forcing of !1.1$1.0 W m\ is in good agreement with a previous estimates of !1.1 W m\ NH average and 0 to !0.5 W m\ forcing over India in July. It follows the same seasonal trend as the sulphate burden. Calculated sulphate aerosol optical depths of 0.01}0.06 agree with previous GCM estimates for the region. Extensive measurements of sulphate size distributions are needed in India for validation of these results. This simple model is in reasonable agreement with detailed GCM results giving us a simple tool to make preliminary estimates of sulphate burdens and direct radiative forcing. We see the following needs in model re"nement and atmospheric measurements as important to this problem: E Re"nement of the chemistry scheme to include seasonal variations in oxidant availability. Oxidant Measurements in air and precipitation are needed. E Extensive measurements of aerosol sulphate concentrations and size distributions and of precipitation composition for region speci"c sulphate budgets, sulphate lifetime and model validation. E Explicit cloud chemistry calculations including resulting sulphate aerosol size distribution. E Mie scattering calculations for sulphate mass scattering coe$cients for typical sulphate aerosol size distributions from cloud-phase reactions. E Improvement of the sulphur dioxide emission inventory for India with detailed technology-linked emission factors for di!erent industrial sectors. E Intercomparison with detailed GCM estimates of regional sulphate burdens and forcings.

Acknowledgements We thank Dr A.P. Mitra (Indian coordinator, INDOEX, NPL) for encouraging our participation in INDOEX. Helpful suggestions from Prof. Henning Rodhe (SU, Sweden) and Dr Olivier Boucher (LOA, France) are much appreciated. We thank Pramod Kulkarni (IIT-Bombay) for "nal calculations, and acknowledge one quarter's salary to PK from AP's contract with The Centre for Integrated Study of Human Dimensions of Global Environmental Change, Carnegie Mellon University, USA.


C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235

Appendix A


A.1. Reaction with hydroxyl radicals (Meng and Seinfeld, 1994) dS(IV) "k [OH]"K !  -& [S(IV)]dt

Andreae, M.O., Browell, E.V., Garstang, M., Gregory, G.L., Harriss, R.C., Hill, G.F., Jacob, D.J., Pereira, M.C., Sachse, G.W., Setzer, A.W., Silva Dias, P.L., Talbot, R.W., Torres, A.L., Wofsy, S.C., 1988. Biomass burning emissions and associated haze layers over Amazonia. Journal of Geophysical Research 93, 1509}1527. Boucher, O., 1997. Personal communication. Boucher, O., Anderson, T.L., 1995. GCM assessment of the sensitivity of direct climate forcing by anthropogenic sulphate aerosols to aerosol size and chemistry. Journal of Geophysical Research 100, 26117}26134. Charlson, R.J., Langner, J., Rodhe, H., Leovy, C.B., Warren, S.G., 1991. Perturbation of the Northern hemisphere radiative balance by backscattering from anthropogenic sulphate aerosols. Tellus 43AB, 152}163. Charlson, R.J., Schwartz, S.E., Hales, J.M., Cess, R.D., Coakley, J.A., Hansen, J.E., Hofmann, D.J., 1992. Climate forcing by anthropogenic aerosols. Science 255, 423}430. Chin, M., Jacob, D.J., Gardner, G.M., Foreman-Fowler, M.S., Spiro, P.A., Savoie, D.L., 1996a. A global three-dimensional model of tropospheric sulphate. Journal of Geophysical Research 101(D13), 18667}18690. Chin, M., Jacob, D.J., 1996b. Anthropogenic and natural contributions to tropospheric sulphates: A global model analysis. Journal of Geophysical Research 101(D13), 18691}18699. Cukier, R.I., Fortuin, C.M., Skuler, K.E., Petschech, A.G., Schaibly, J.H., 1973. Study of the sensitivity of coupled reaction systems to uncertainties in rate coe$cients. Journal of Chemical Physics 159(8), 3873}3880. Feichter, J., Kjellstrom, E., Rodhe, H., Dentener, F., Lelieveld, J., Roelofs, G., 1996. Simulation of the tropospheric sulphur cycle in a global climate model. Atmospheric Environment 30, 1693}1707. Galloway, J.N., Savoie, D.L., Keene, W.C., Prospero, J.M., 1993. The temporal and spatial vairability of scavenging ratios for nss sulfate, nitrate, methanosulfate and sodium in the atmosphere over the north Atlantic ocean. Atmospheric Environment 27A, 235}250. Hastenrath, S., 1985. Climate and Circulation of the Tropics, D. Reidel Publishing Company, Dordrecht, Holland, pp. 25}36. Ibusuki, T.M., Takeuchi, K., 1987. Sulphur dioxide oxidation by oxygen catalysed by mixtures of Manganese (II) and Iron (III) in aqueous reactions at environmental reaction conditions. Atmospheric Environment 21, 1555}1560. Jayaraman, A., Lubin, D., Ramachandran, S., Ramanathan, V., Woodbridge, E., Collins, W.D., Zalpuri, K.S., 1998. Direct observations of aerosol radiative forcing over the tropical Indian ocean During the Jan}Feb. 1996 Pre-INDOEX Cruise. Journal of Geophysical Research 103, 13827}13836. Kaul, O.N., 1993. Forest biomass burning in India. In: Achanta, A.N. (Ed.), Global Climate Change: An Indian Perspective. Tata Energy Research Institute, New Delhi, pp. 91}110. Kaul, O.N., Shah, V.N., 1993. Grassland biomass burning in India. In: Achanta, A.M. (Ed.), Global Climate Change: An Indian Perspective. Tata Energy Research Institute, New Delhi, pp. 65}90. Khemani, L.T., Momin, G.A., Naik, M.S., Vijaykumar, R., Ramanamurthy, B.H.V., 1982. Chemical composition and

k "1.2;10\ molecules\ cm s\.  A.2. Reaction with oxygen (Ibusuki and Takeuchi, 1987) For pH (4.2,

d[S(IV)] 8432 ! "K exp ! [Fe>][Mn>]  S(IV)dt ¹ [H>]\ "K . - For pH '4.2, 8289 d[S(IV)] "K exp ! [Fe>][Mn>] !  [S(IV)]dt ¹ [H>] "KO


!8432 M\ s\ ¹

K " 8.1;10 exp 

K " 5.47;10 exp 

HSO "1.23exp !3020 

!8432 ¹

M\ s\

1 1 ! ¹ 298

M\ atm\.

A.3. Reaction with ozone (Maahs, 1983)


d[S(IV)] k ! " k #  HO [O ]"KO     [S(IV)]dt H> 996 k "2.56;10 exp ! , M\ s\  ¹ 4131 k "4.39;10 exp ! , M\ s\  ¹ HO "11.1;10\ exp !2300 

1 1 ! ¹ 298


M\ atm\. A.4. Reaction with hydrogen peroxide (Martin and Damschen, 1981) dS(IV) k HH O [H O ] ! "      E [S(IV)]dt 0.1#[H>] "KH O



HH O "7.1;10 exp  

6957 , M\ atm\. ¹

1 1 k "8;10 exp !3650 !  ¹ 298

, M\ s\

C. Venkataraman et al. / Atmospheric Environment 33 (1999) 3225}3235 size distribution of atmospheric aerosols over the Deccan plateau, India. Tellus 34, 151}158. Khemani, L.T., 1989. Physical and chemical characteristics of atmospheric aerosols. In: Cheremisino!, P. (Ed.), Encyclopaedia of Environmental Control Technology. Vol. 2, Gulf Publishing Company, Houston, pp. 401}453. Khemani, L.T., Momin, G.A., Rao, P.S.P., Vijaykumar, R., Safai, P.D., 1995. Study of surface ozone behaviour at urban and forested sites in India. Atmospheric Environment 29(16), 2021}2024. Kiehl, J.T., Briegleb, B.P., 1993. The relative roles of sulfate aerosols and greenhouse gases in climate forcing. Science 260, 311}314. Kiehl, J.T., Rodhe, H., 1995. Modelling geographical and seasonal forcing due to aerosols. In: Charlson, R.J., Heintzenberg, J. (Eds.), Aerosol Forcing of Climate. Wiley, Chicester, UK, pp. 281}297. Krishna Moorthy, K., Nair, P.R., Prasad, B.S.N., Muralikrishnan, N., Gayathri, H.B., Narasimha Moorthy, B., 1993. Results from the MWR Network of IMAP. Indian Journal of Radio and Space Physics 22, 243}258. Langner, J., Rodhe, H., 1991. A global three dimensional model of the tropospheric sulfur cycle. Journal of Atmospheric Chemistry 13, 225}263. Lelieveld, J., Crutzen, P.J., Rodhe, H., 1989. Zonal average cloud characteristics for global atmospheric chemistry modelling. Report CM-76, Department of Meteorology, University of Stockholm. Lelieveld, J., Heintzenberg, J., 1992. Sulphate cooling e!ect on climate through in}cloud oxidation of anthropogenic SO .  Science 258, 117}120. Maahs, H.G., 1983. Kinetics and mechanisms of oxidation of S(IV). by ozone in aqueous solution with particular reference to SO conversion in nonurban tropospheric clouds. Journal  of Geophysical Research 88, 10721}10732. Martin, L.R., Damschen, D.E., 1981. Aqueous oxidation of sulfur dioxide by hydrogen peroxide at low pH. Atmospheric Environment 15, 1615}1626. Martin, L.R., Good, T.W., 1991. Catalysed oxidation of sulfur dioxide in solutions: the iron}manganese synergism. Atmospheric Environment 25A, 2395}2399. Mazumdar, B.K., Chakravorty, M., Mukherjee, D.K., 1985. Problems and prospects of coal utilisation in India. Fuel Sci. Technol. 4(1), 15}35. McRae, G.J., Tilden, J.W., Seinfeld, J.H., 1982. Global sensitivity analysis: a computational implementation of the Fourier


Amplitude Sensitivity Test (FAST). Computers and Chemical Engineering 6, 15}25. Meng Z., Seinfeld, J.H., 1994. On the source of the submicrometer droplet mode of urban and regional aerosols. Aerosol Science and Technology 20, 253}265. Middleton, P., Kiang, C.S., Mohnen,V.A., 1980. Sulfur dioxide oxidation; a theoretical estimate of the relative contributions of di!erent pathways. Atmospheric Environment 14, 463}472. Mitra, A.P., 1992. Greenhouse gas emissions in India } 1992 update. Technical Report, Council for Scienti"c and Industrial Research and Ministry of Environment and Forests, New Delhi. Parashar, D.C., Granat, L., Kulshreshta, U.C., Pillai, A.G., Naik, M.S., Momin, G.A., 1996. Chemical composition of precipitation in India and Nepal: a preliminary report on an Indo}Swedish project on atmospheric chemistry. Report CM90, Department of Meteorology, University of Stockholm. Penner, J.E., Dickinson R.E., O'Neill, C.A., 1992. E!ects of aerosols from biomass burning on the global radiation budget. Science 256, 1432}1434. Pham, M., Muller, J.-F., Brasseur, G.P., Granier, C., Megie, G., 1995. A three dimensional study of the tropospheric sulphur cycle. Journal of Geophysical Research 100(D12), 26061}26092. Pham, M., Muller, J.-F., Brasseur, G.P., Granier, C., Megie, G., 1996. A three-dimensional study of the tropospheric sulphur cycle. Atmospheric Environment 30(11), 1815}1822. Pilinis, C., Pandis, S.N., Seinfeld, J.H., 1995. Sensitivity of direct climate forcing by atmospheric aerosols to aerosol size and composition. Journal of Geophysical Research 100(D9), 18739}18754. Savoie, D.L., Prospero, J.M., Nees, R.T., 1987. Nitrate, nss sulfate and mineral aerosols over the northwestern Indian ocean. Journal of Geophysical Research 92, 933}942. Savoie, D.L., Prospero, J.M., Nees, R.T., 1987. Washout ratios of nitrate, nss sulfate, and sea-salt on Virginia Key, Florida and on American Samoa, Atmospheric Environment 21, 103}112. Sinha, C.S., Joshi, V., 1997. Biofuel demand estimation in the rural domestic energy sector of India. In: Venkata Ramana, P. (Ed.). Rural and Renewable Energy: Perspective from Developing Countries, Tata Energy Research Institute, New Delhi. Spiro, P.A., Jacob D.A., Logan, J.A., 1992. Global inventory of sulphur emissions with 13;13 resolution. Journal of Geophysical Research 97 (D5) 6023}6036.