Acta Geophysica vol. 55, no. 3, pp. 383-397 DOI 10.2478/s11600-007-0009-3
Effect of synoptic meteorological conditions on aerosol properties over urban environment: A study over tropical urban region of Hyderabad, India Shailesh Kumar KHAROL and K.V.S. BADARINATH Atmospheric Science Section, National Remote Sensing Agency Department of Space, Government of India, Balanagar Hyderabad – 500 037, India; e-mail:
[email protected]
Abstract Aerosol particle size distribution and chemical properties are important in studies related to human health and climate. The present study describes an analysis of aerosol mass loading, Aerosol Optical Depth (AOD), black carbon aerosol mass concentration and carbon monoxide over tropical urban region of Hyderabad, India, during March 2006, coinciding with active forest fires season over India. Aerosol optical depth, particulate matter mass loading and carbon monoxide were observed to be high on days with air mass coming from north of the study area. Spatial occurrence of forest fires was analysed using MODIS daytime data and DMSP-OLS nighttime data sets. Aerosol optical depth measured using Microtops–II sunphotometer correlated well with MODIS derived AOD values. Results of the study suggested that synoptic meteorological conditions play an important role in the observed aerosol properties over the study area during the forest fire season. Key words: aerosols, black carbon, particulate matter, biomass burning, urban environment, forest fire. 1. INTRODUCTION
The aerosol mass size distribution and its temporal and spatial variability are fundamental aerosol property. The number-size distribution and hygroscopic properties of the sub micrometer aerosol particles are thus crucial properties for the life cycle of the atmospheric aerosol, as well as the radiative forcing of aerosols on regional and global © 2007 Institute of Geophysics, Polish Academy of Sciences
384
S.K. KHAROL and K.V.S. BADARINATH
climate. Airborne particulate matter is the visible element of air pollution with very diverse effects, ranging from human health to climate forcing. Fine particles have intrinsically low sedimentation and deposition velocities, and their atmospheric residence time is of the order of days. Atmospheric Particles are often characterized by their size cutoff, e.g. as PM1.0 , PM2.5 and PM10 . Particulate Matter (PM) includes acids, metals, and the solid or liquid droplets suspended in air, and other harmful airborne substances that can be breathed into the lungs. PM sources include automobiles, diesel engines, power plants, industrial facilities, wood combustion, and dust from deserts/roads. Some particles are large or black enough to be seen as soot or smoke. Airborne particulate matter 2.5 microns or less in size (PM2.5) comprises a complex mixture of particles composed of sulfate, nitrate, ammonium, organic carbon, elemental carbon, and organic and inorganic compounds. Since the properties and behavior of particles are highly dependent on the particle size, the particle size distribution can provide detailed information on mode distributions and also give evidence on the formation and transformation of particles in the atmosphere (Whitby 1978, Hering et al. 1997). These atmospheric particles can be anthropogenic or natural in origin. Both anthropogenic and natural particulate material can occur from either primary or secondary processes. Particles also have a complicated growth mechanism of condensation, coagulation, and adsorption. Severe health effects are associated with exposure to excess levels of airborne fine particulate matter (PM2.5). Atmospheric aerosol particles in urban areas cause decrease of visibility (e.g., Finlayson-Pitts et al. 2000) and health effects (Dockery and Pope 1994). A significant number of health problems related to atmospheric aerosols is believed to be due to particles having diameters less than 1.0 µm, because these particles can penetrate deep into the respiratory system (Dockery and Pope 1994). Black carbon (BC) aerosol is the major anthropogenic component of atmospheric aerosol system which has significantly different optical and radiative properties as compared to the other normal constituents. Human activities affect the aerosol particle concentration levels and so the black carbon aerosol concentration is a good tracer of anthropogenic activities. BC, due to its large absorption over a wide wavelength range, could offset significantly or even reverse the cooling effect, due to aerosol scattering (Schwartz 1996, Haywood and Shine 1997). Burning of biomass and fossil fuels, automobile exhaust, aircraft emissions and forest fires are the major sources of BC. Due to its environmental and climate significances, as well as anthropogenic nature of its origin, characterization of BC is attracting considerable interest in the recent years (Hansen et al. 2000). The present study provides a comprehensive account of temporal variation of Aerosol Optical Depth (AOD), aerosol mass size distribution, black carbon aerosol mass concentration and carbon monoxide over tropical urban environment of Hyderabad, India, during forest fire season using satellite data and ground based measurements.
EFFECT OF METEOROLOGICAL CONDITIONS ON AEROSOL PROPERTIES
385
2. STUDY AREA
The study area of Hyderabad is located between 17°10΄ and 17°50΄N latitude and 78°10΄ and 78°50΄E longitude. Hyderabad is the fifth largest city in India; its population is 3,449,878 inhabitants according to the census of 2001, and it is a purely urbanized area. The climate of the region is semi-arid type with total rainfall amount of ~700 mm occurring mostly during monsoon season in the period June–October. The minimum and maximum temperatures during March 2006 were 17° and 36°C, respectively, under clear sky conditions. The relative humidity values in March are normally high during nighttime (95%), and are in the range of 25 to 45% during daytime. All the synchronous measurements for the study were carried out under clear sky conditions in the premises of the National Remote Sensing Agency (NRSA) campus at Balanagar (17°28΄N and 78°26΄E) located well within the urban center. 3. DATA SETS AND METHODOLOGY
Continuous measurements of particulate matter (PM) size distribution were performed with GRIMM aerosol spectrometer model 1-108 (Le Canut et al. 1996). The GRIMM instrument works on the principle of counting the number of particles as it crosses a focused laser beam. Scattering induced by particles of various size is measured by a photo-diode detector, amplified and finally binned to give the distribution of particulate matter in 15 different size classes from 0.30 to 20 µm. The instrument is capable of counting particles from 1 particle dm-3 of air to 2 million particles dm-3 and the lower detectable mass is 0.1 µg/m3. In addition, continuous and near-real-time measurements of the mass concentration of aerosol BC were carried out for March 2006, using an aethalometer, model AE-21 of Magee Scientific, USA (Cooke et al. 1997, Borak et al. 2003). The instrument draws ambient air from an altitude of 3 m above the ground using its inlet tube and its pump from the same location as that of GRIMM spectrometer. The BC mass concentration is estimated by measuring the change in the transmittance of a quartz filter tape, onto which the particles impinge. The instrument was operated at a time base of 5 minutes, round the clock with a flow rate of 3 LPM. The instrument was factory calibrated and errors in the measurements of BC are ca. ±2%. Synchronous and continuous observations of Aerosol Optical Depth (AOD) were carried out using a handheld multichannel sun-photometer (Microtops-II, Solar Light Co., USA) at six wavelength bands centered around 380, 440, 500, 675, 870 and 1020 nm (Leckner 1978). The sun-photometer works on the principle of measuring the surfacereaching solar radiation intensity at the specified wavelength bands and converts to optical depth using Langley method. The AOD τA(λ) was retrieved from the measuring data by accounting for Rayleigh scattering τR(λ) and the contribution of gas absorbers: τ A (λ ) = τ (λ ) − τ R (λ ) − τ O 3 (λ ) − τ H 20 (λ ) . The Rayleigh scattering has been calculated by formula τ R (λ ) = (P/P0)×0.008735×λ-4.08.
386
S.K. KHAROL and K.V.S. BADARINATH
In this formula, P is the actual air pressure in hPa and P0 = 1013.25 hPa. Continuous measurements of carbon monoxide (CO) were carried out using CO11M analyser of Environment SA, France. The meteorological data during March 2006 were collected from the METOS weather station installed at the measurement site. The hourly recorded air temperature, relative humidity, wind speed and wind direction data were analysed for this study. Daily data sets of DMSP-OLS for February 2006 were processed for generating nighttime fire products over the Indian region. DMSP operates in sun-synchronous orbits with nighttime overpasses ranging from about 7 pm to 10 pm local time with a swath width of 3000 km. The OLS is an oscillating scan radiometer with two spectral bands. The visible band pass straddles the visible and near-infrared portion of the spectrum (0.5 to 0.9 µm) and the thermal band pass covers the 10.5 to 12.5 µm region (Elvidge et al. 1997). The low light sensing capabilities of the OLS at night permit the measurement of radiances down to 10-9 watts/cm2/sr. Fires present at the Earth’s surface at the time of the nighttime overpass of the DMSP are readily detected in the visible band data. In addition, daytime MODIS data over the region were processed for forest fires using the thermal bands data (Giglio et al. 2003). The Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model of NOAA Air Resources Laboratory was used to study the origin of air masses over the study area (Espozito et al. 2004). MODIS level 2 daily aerosol optical depth products from Terra (MOD04 V004) website at three wavelengths 440, 550 and 660 nm were analysed for March 2006. 4. RESULTS AND DISCUSSION
Continuous measurements of Particulate Matter PM1.0, PM2.5 and PM10 aerosols were carried out using GRIMM aerosol spectrometer and are shown in Fig. 1 for March 2006. The average daily PM1.0, PM2.5 and PM10 concentrations are in the range of 4.23 to 26.96 µg/m3, 6.04 to 42.38 µg/m3 and 6.87 to 45.34 µg/m3. PM2.5 values were above the EPA standards (i.e., > 65 µg/m3) on days with higher aerosol mass loading during last week of March 2006 corresponding to Julian days 86–90 (Kousa et al. 2002). Figure 2 shows the Julian day variation of average PM1.0 aerosol particulate matter values over the study area. The higher concentration of PM1.0 values on Julian days 69–72 and 78–85 suggests accumulation mode particle loading over the urban region, and the primary sources for these particles are vehicular traffic and other anthropogenic disturbances. The MODIS satellite data and the nighttime satellite data sets from DMSPOLS were processed for forest fires over the Indian region for identifying possible sources for such a high accumulation particle loading (Kharol and Badarinath 2006). Figure 3 depicts the fifteen-day time composite of forest fires over the region for March 2006 with measurement site of Hyderabad depicted on the figure. MODIS derived active fire locations overlaid on NOAA-AVHRR false color composite (FCC) of 22 March 2006 are shown in Fig. 4.
EFFECT OF METEOROLOGICAL CONDITIONS ON AEROSOL PROPERTIES
387
PM1.0 (µg/m3)
60
40
20
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
72
74
76
78
80
82
84
86
88
90
0
80
PM2.5 (µg/m3)
60
40
20
0
100
PM10 (µg/m3)
80 60
40 20
70
68
66
64
62
60
0
JULIAN DAY
Fig. 1. Variations of aerosol particulate matter PM1.0, PM2.5, PM10 (in µg/m3) during March 2006.
The total fire counts derived from DMSP-OLS nighttime data over Indian region and also state wise fire counts during March 2006 are shown in Figs. 5 and 6. It can be seen from Fig. 5 that there is an increased fire activity over the region during the last week of March, which is reflected in higher AOD values (Fig. 8) and particulate matter loading (Figs. 1 and 2), and black carbon aerosol concentration (Fig. 11) over
S.K. KHAROL and K.V.S. BADARINATH
388 30.00
c
a
20.00
10.00
90
88
86
84
82
80
78
76
72
68
66
64
62
60
74
e
d
0.00
70
PM1.0 (µg/m3)
b
JULIAN DAY
Fig. 2. Julian day variation of PM1.0 during March 2006.
Fig. 3. The DMSP-OLS nighttime fifteen-day time composite of the forest fire locations over Indian region during March 2006: Julian days 60-74 (left), Julian days 75-90 (right).
measurement site of Hyderabad. The state wise fire counts derived from DMSP-OLS nighttime data (Fig. 6) and MODIS daytime fires (Fig. 4) suggested higher incidence of fires in states like Chattisgarh, Orissa, Andhra Pradesh and Maharashtra, that are spatially nearer to the measurement site of Hyderabad. The HYSPLIT model wind back trajectory analyses on days with high and low values of PM1.0 of Fig. 2 are shown in Fig. 7. Figure 7 shows the back trajectories for different dates at 500, 1000 and 1500 m altitude for a given atmospheric level taken from the NOAA HYSPLIT model for the measurements site. It can be seen from Figs. 7a, b, c that the airmass movement over the tropical urban area of Hyderabad was
EFFECT OF METEOROLOGICAL CONDITIONS ON AEROSOL PROPERTIES
389
Fig. 4. MODIS active fire locations during Julian days 79th to 90th, 2006 overlaid on Julian day 81, 2006 NOAA AVHRR False Color Composite (FCC).
NUMBER OF FIRE PIXELS
5000
4000
3000
2000
90
88
86
84
82
80
78
76
74
72
70
68
66
64
62
0
60
1000
JULIAN DAY
Fig. 5. Variation of Forest Fire counts derived from DMSP-OLS nighttime data over Indian Region for March 2006.
from northern region of India having forest fires during that period, which is expected to bring the polluted airmass directly over the urban area. The higher values of PM1.0 coincided with air mass coming from north of the study area having forest fires and
S.K. KHAROL and K.V.S. BADARINATH
390
NUMBER OF FIRE PIXELS
7000 6000 5000 4000 3000 2000
Kerala
Goa
Tamilnadu
AP
Karnataka
Orissa
Maharastra
Tripura
Chattisgarh
Gujarat
Mizoram
Manipur
Jharkhand
MP
Meghalaya
Bihar
W Bengal
Sikkim
Nagaland
Assam
Arunachal
UP
Rajasthan
Haryana
Uttaranchal
HP
Punjab
0
J&K
1000
STATE
Fig. 6. Forest fire counts derived from DMSP-OLS nighttime data in different states of Indian region during March 2006.
lower values of PM1.0 correlated with air mass coming from south of the study area. Figures 7d, e show the airmass flow from the Bay of Bengal oceanic region to the measurement site location. These air masses having increased sea salt component cannot travel longer distance due to low wind speeds and hence the fine aerosol concentration is not significantly increased (Niranjan et al. 2005). Figure 8 shows the Aerosol Optical Depth (AOD) variation at 500 nm as a function of Julian day for March 2006 measured using MICROTOPS-II sunphotometer. The AOD values vary between 0.3 to 1; higher AOD were observed on days with forest fires occurring to the north of the measurement site. The NOAA-HYSPLIT model runs suggested that on higher aerosol mass loading days, winds are coming from north of the study area (Figs. 7a, b, c). Ångström turbidity coefficient β estimated from wavelength variation of aerosol optical depth suggested higher values during the last week of March 2006 suggesting turbid conditions with higher accumulation mode particle loading. The values of Ångström turbidity coefficient β are of the order of 0.18 on normal day (Julian day 74) and 0.32 during high aerosol loading day (Julian day 81). Figure 9 shows the variation of aerosol optical depth (AOD) at 500 nm on normal day (Julian day 74) and high aerosol loading day (Julian day 81). It can be noticed from Fig. 9 that AOD values increased by ~30% on Julian day 81 due to biomass burning activities in the region. TERRA MODIS daytime derived mean aerosol optical depth (AOD) at 550 nm for March 2006, shown in Fig. 10, suggested higher AOD over the region coinciding with the occurrence of forest fires (Fig. 3). Figure 11 shows the diurnal variation of black carbon (BC) aerosol for March 2006 measured using aethalometer. There was rainfall during Julian days 62–70 caus-
EFFECT OF METEOROLOGICAL CONDITIONS ON AEROSOL PROPERTIES
(a)
(b)
391
(c)
(d)
(e)
Fig. 7. HYSPLIT Model Black trajectory plots on different days corresponding to high and low PM1.0 conditions shown in Fig. 2. MARCH 2006
TIME (hrs)
15 13 11 9 60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
JULIAN DAY
Fig. 8. Julian day variation of Aerosol Optical Depth at 500 nm during March 2006.
90
S.K. KHAROL and K.V.S. BADARINATH
392 1.000
74 JULIAN DAY 81 JULIAN DAY
AOD (500nm)
0.800 0.600 0.400 0.200 0.000 9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
TIME (hrs)
Fig. 9. Variations in Aerosol Optical Depth at 500 nm on normal day (rhombs – 74th Julian day) and high aerosol loading day (triangles – 81st Julian day).
Fig. 10. TERRA-MODIS daytime mean Aerosol Optical Depth at 550 nm for March 2006.
ing scavenging of BC as observed in Fig. 10. BC variations showed gradual build up during early morning hours and a sharp peak occurring between 6:00 and 9:00 LT almost an hour after the local sunrise. This arises from the combined effects of (i) the well-known fumigation effect in the boundary layer, which brings-in aerosols from the nocturnal residual layer shortly after the sunrise (Stull 1988, Babu et al. 2002) and (ii) the morning build up of local anthropogenic activities in the urban area
EFFECT OF METEOROLOGICAL CONDITIONS ON AEROSOL PROPERTIES
393
MARCH 2006
22 20 18 16
TIME (hrs)
14 12 10 8 6 4 2 0 60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
JULIAN DAY
Fig. 11. Julian day variations of the black carbon (BC) aerosol mass concentration during March 2006.
from where the wind is still directed. BC values were observed to be high during the last week of March 2006 and have been attributed to forest fires towards north of the study area over the region. BC values four times higher than normal values on these days are related to active forest fires over the region. It can be seen from Fig. 12 that both carbon monoxide and black carbon aerosol mass concentrations were high during the last week of March compared to normal day. Such an increase in both CO and BC could result from burning activities and the occurrence of forest fires (Fig. 3) and favorable wind conditions from north of the study area (Figs. 7b, c) leading to higher loading of aerosols over the study area. BC mass concentration and PM variations with meteorological parameters suggested positive correlation with wind direction and temperature and negative correlation with relative humidity (correl. coeff = 0.61). BC mass concentrations showed a good correlation with PM1.0 (correl. coeff = 0.92). Scatter plot of MODIS derived AOD with sunphotometer measurements at different wavelengths shown in Fig. 13. The good correlation between satellite derived AOD with ground measurements suggests possible use of satellite derived AOD for inferring spatial and temporal dynamics of aerosol properties.
0.4
12
0.35
10
0.3 0.25
8 6
0.2 0.15
4
0.1
2
0.05 0
3
BC (ug/m )
14
0 3
5
7
9
11
13
15
17
19
21
120
4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
100 3
BC (ug/m )
23
80 60 40 20 0 1
3
5
7
9
11
13
15
HOUR OF DAY
17
19
21
23
CO (ppm)
1
C O (ppm)
S.K. KHAROL and K.V.S. BADARINATH
394
BC CO
Fig. 12. Diurnal variation of black carbon (BC) aerosol and carbon monoxide (CO) on normal day (74th Julian day – top) and high aerosol loading day (81st Julian day – down). 5. CONCLUSIONS
In this study, the impact of the forest fires on the aerosol mass size distribution, black carbon (BC) aerosol mass concentration, carbon monoxide (CO) and aerosol optical depth (AOD) were analysed over urban region of Hyderabad, India. The results of the analysis suggested that: – Aerosol optical depth values, particulate matter loading, black carbon aerosol mass concentration and carbon monoxide showed higher values during periods with wind direction trajectories from north of the study area. – MODIS day time and DMSP-OLS nighttime data analysis suggested increased incidence of forest fires over the region during later part of March 2006. Air mass trajectories coming from north of the study area resulted in increased aerosol optical depth (AOD), black carbon (BC) aerosol concentration, aerosol particle mass size distribution and carbon monoxide (CO) values. – MODIS measured aerosol optical depth at 470, 550 and 660 nm correlated reasonably well with sunphotometer based ground measurements.
EFFECT OF METEOROLOGICAL CONDITIONS ON AEROSOL PROPERTIES MARCH 2006
0.8
MODIS AOD (470nm)
395
y = 0.5532x - 0.0142 R2 = 0.6264
0.6 0.4
0.2
(440nm) 0 0.4
0.2
0.6
0.8
1
1.2
MODIS AOD (550nm)
1 y = 0.6146x - 0.0316 R2 = 0.37
0.8 0.6 0.4 0.2
(500nm)
0 0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
MODIS AOD (660nm)
0.8 y = 0.6727x + 0.0739 R2 = 0.4548
0.6
0.4 0.2
(675nm) 0 0.1
0.2
0.3
0.4
0.5
0.6
0.7
MICROTOPS AOD
Fig. 13. Comparison of MODIS derived Aerosol Optical Depth (AOD) at 470, 550 and 660 nm with sunphotometer measurements.
– Results of the study indicate that emissions from forest fires over Indian region influence urban air quality of Hyderabad during fire season. – The study is the first of this kind over the region to integrate satellite data derived fire occurrences and synoptic meteorological conditions to explain the observed variations in aerosol properties. A c k n o w l e d g e m e n t s . The authors thank Director, NRSA and Dy. Director (RS&GIS-AA) for necessary help at various stages, and ISRO-GBP for funding support. The authors thank anonymous referees for comments and suggestions.
396
S.K. KHAROL and K.V.S. BADARINATH
References Babu, S.S., and K. Krishna Moorthy, 2002, Aerosol black carbon over a tropical coastal station in India, Geophys. Res. Lett. 29, 23, 2098, doi: 10.1029/2002GL015662. Borak, J., G. Sirianni, H.J. Cohen, S. Chemerynski and R. Wheeler, 2003, Comparison of NIOSH 5040 method versus aethalometery to monitor diesel particulate in school buses and at work sites, AIHA Journal 64, March/April, 260-268. Cooke, W.F., S.G. Jennings and T.G. Spain, 1997, Black carbon measurements at Mace Head, 1989 –1996, J. Geophys. Res. 102, 25,339-25,346. Dockery, D.W., and C. Pope, 1994, Acute respiratory effects of particulate air pollution, Ann. Rev. Public Health 15, 107-132. Elvidge, C.D., K.E. Baugh, E.A. Kihn, H.W. Kroehl and E.R. Davis, 1997, Mapping of city lights using DMSP Operational Linescan System data, Photogram Eng. Remote Sensing 63, 727-734. Espozito, F., L. Leone, G. Pavese, R. Restieri and C. Serio, 2004, Seasonal variation of aerosols properties in South Italy: A study on aerosol optical depths Ångström turbidity parameters and aerosol size distributions, Atmos. Environ. 38, 1605-1614. Finlayson-Pitts, B.J., and J.N. Pitts Jr., 2000, Chemistry of the Upper and Lower Atmosphere, Academic Press, San Diego. Giglio, L., J. Descloitres, C.O. Justice and Y.J. Kaufman, 2003, An enhanced contextual fire detection algorithm for MODIS, Remote Sensing Environ. 87, 273-282. Hansen, J., M. Sato, R. Reudy, A. Lacis and V. Oinas, 2000, Global warming in the 21st century: An alternative scenario, Proc. Nat. Acad. Sci. 97, 9875-9880. Haywood, J.M., and K.P. Shine, 1997, Multi-spectral calculations of the radiative forcing of tropospheric sulphate and soot aerosols using a column model, Q. J. R. Meteorol. Soc. 123, 1907-1930. Hering, S., A. Eldering and J.H. Seinfeld, 1997, Bimodal character of accumulation mode particle mass distributions in southern California, Atmos. Environ. 31, 1-11. Kharol, S.K., and K.V.S. Badarinath, 2006, Impact of biomass burning on aerosol properties over tropical urban region of Hyderabad, India, Geophys. Res. Lett. 33, L20801, doi: 10.1029/2006GL026759. Kousa, A., L. Oglesby, K. Koistinena, N. Künzlib and M. Jantunena, 2002, Exposure chain of urban air PM2.5––associations between ambient fixed site, residential outdoor, indoor, workplace and personal exposures in four European cities in the EXPOLISstudy, Atmos. Environ. 36, 18, 3031-3039. Le Canut, P., M.O. Andreae, G.W. Harris, F.G. Wienhold and T. Zenker, 1996, Airborne studies of emissions from savanna fires in southern Africa: 1. Aerosol emissions measured with a laser optical counter, J. Geophys. Res. 101, D19, 23,615-23,630. Leckner, B., 1978, The spectral distribution of solar radiation at the Earth’s surface – Elements of model, Sol. Energy 20, 143-150. Niranjan, K., B. Melleswara Rao, P.S. Brahmanandam, B.L. Madhavan, V. Sreekanth and K. Krishna Moorthy, 2005, Spatial characteristics of aerosol physical properties over the northeastern parts of peninsular India, Ann. Geophysicae 23, 3219-3227.
EFFECT OF METEOROLOGICAL CONDITIONS ON AEROSOL PROPERTIES
397
Schwartz, S.E., 1996, The white house effect – Short wave radiative forcing of climate by anthropogenic aerosols – An overview, J. Aerosol Sci. 27, 359-382. Stull, R.B., 1988, An Introduction to Boundary Layer Meteorology, Kluwer Academic Publ., Dordrecht. Whitby, K.T., 1978, The physical characteristics of sulfur aerosols, Atmos. Environ. 12, 135159. Received 15 June 2006 Accepted in revised form 14 December 2006
