Geophysical Research Letters

Latitudinal distribution of aerosol black carbon and its mass fraction to composite aerosols over peninsular India during winter season

Authors


Abstract

[1] During a land campaign to characterise the spatial distribution of aerosols over peninsular India during the winter season, extensive, collocated, and spatially resolved measurements of mass concentration of the composite aerosols (MT) as well as that (MB) of aerosol Black Carbon (BC) were made over different environments (coastal, industrial, urban, village, remote, semiarid) of the western peninsular India. High concentrations of BC, >2.5 μg m−3, were observed along the west coast, from ∼8°N up to 14.5°N, and moderate values (1.0 to 2.5 μg m−3) over inland regions from 15 to 18°N. Latitudinally, BC concentration decreased from south to north, @∼160 ng m−3 for every degree increase in latitude. The spatial pattern of BC mass fraction differed from that of MB, with regions of high (8 to 16%) ratios spreading more interior, implying higher fractional load of BC at locations where the BC concentrations remain lower.

1. Introduction

[2] Aerosol black carbon (BC or soot) is the principal light-absorbing component of atmospheric aerosols, having strong absorption over a wide wavelength range from UV to IR, and contributing significantly to aerosol radiative forcing [Jacobson, 2001; Babu et al., 2004]. Being a by-product of all low temperature/incomplete combustion processes, BC particles are generally in the sub-micron size regime and are considered as tracers of anthropogenic impact on environment [Cachier, 1998]. Significant amount of BC over reflecting land surfaces or over highly scattering clouds enhances atmospheric absorption, and at times, even reverse the ‘white house effect’ due to scattering aerosols. Not only the absolute mass concentration; but its share to the composite aerosols as well as the altitude variation are also important in estimating the radiative and environmental impact of BC [Babu et al., 2004]. Thus the study of BC has assumed great topical importance.

[3] However, systematic and extensive observations leading to the characterization of BC over Indian landmass are limited except by Babu et al. [2002] from Bangalore, Babu and Moorthy [2002] from Trivandrum, and Tripathi et al. [2005] from Kanpur. These fixed station measurements, some of them using long-term data base, though have helped to understand the temporal changes in BC characteristics associated with natural changes in atmospheric circulation patterns at these locations, understanding of the spatial variation is very limited. In view of its high population density, large heterogeneity in human activities, and diverse living habits, such information is needed for the Indian region to assess the potential of various source processes on the one hand and to understand and estimate the consequence on radiative forcing on the other. Such an attempt was made during February 2004 under the ISRO Geosphere Biosphere Program [Moorthy et al., 2005] and the results are presented in the following sections.

2. Measurement Details and Data Base

[4] Mass concentration (MB) of aerosol BC was measured at near real time using two Aethalometers (Models AE 42 &AE 21 of Magee Scientific), while the mass concentration (MT) of the composite aerosols were measured using two ten-channel Quartz Crystal Microbalance (QCM) impactors (PC-2 of California Measurements Inc). Both the aethalometers sampled air from ∼3 m above ground, at a flow rate 4 l min−1 with a time base of 5 min. The measured optical attenuation is converted to MB using an effective absorption coefficient of ∼16 m2g−1 [Hansen, 1996; Babu et al., 2004]. Following the details given by Hansen [1996], Babu and Moorthy [2002] and Babu et al. [2004], the uncertainties in the estimates of MB is <∼50 ng m−3 (2 to 5% of the measured values).

[5] Mass concentration (MT) and mass size distribution (mci) of the composite aerosols (over the size range 0.05 to 25 micrometer) were obtained from near simultaneous, but less frequent, measurements using the collocated QCM impactor, operated at a flow rate of 0.24 l min−1 and is sampled for a duration varying between 240 to 360 s depending on the particle concentration at the sampling location. Samples were collected at an interval varying between 30 to 60 min during the data collection period. Following the error budget [Pillai and Moorthy, 2001] the uncertainty in MT is between 15 to 25%; being higher at lower concentrations.

[6] The measurement sites were distributed in latitude from 8.5° to 22.5°N and the longitude from 74 to 79°E. The details of the topography of the region, sampling location and protocols are given in detail in an earlier paper [Moorthy et al., 2005]. Measurements were made independently by two teams, each carrying instruments in separate vehicles along the routes shown in Figure 1, covering the western part of peninsular India. The track with solid circles shows the route covered by Team-1 (Space Physics Laboratory, Trivandrum), and track with closed triangles shows the route covered by Team-2 (National Remote Sensing Agency, Hyderabad). Team-1 started from Trivandrum (8.5°N, 77°E; TVM in Figure 1) on 1 February 2004 and travelled along the west coast till Karwar (KRW, 14.9°N, 74.1°E) before deviating to the inland vegetated forest regions followed by the dry, semi-arid locations until Sholapur, (SLP, 17.97°N, 76°E), from where it moved southwards and reached Shadnagar (SNR, 17.03°N, 78.19°E), while Team-2 covered the inland, semiarid regions from Hyderabad (HYD, 17.3°N, 78.5°E) to Indore (IND, 22.6°N, 75.9°E) and returned (along the same track) to SNR. At Shadnagar an intercomparison of the instruments was made by under identical ambient condition and an agreement was seen within 5%. In the return leg Team-1 made observations towards south and crossed Western Ghats and reached TVM.

Figure 1.

Route map followed by the teams during the campaign. The points on the track lines identify the locations of measurement.

[7] During this campaign, independent measurements were made at 40 sites by the two teams put together following a common protocol [Moorthy et al., 2005]. Accordingly, at each location, the measurements were made during the daytime period between 09:00 LT and 16:00 LT. During this period the daytime atmospheric boundary layer is well evolved [Moorthy et al., 2005] and the diurnal variations in the aerosol concentration were, generally, the lowest [Pillai and Moorthy, 2001; Moorthy et al., 2005]. The clear skies, calm winds, and absence of rainfall during the entire campaign period provided ideal condition for making the spatially resolved measurements. Sampling sites were selected to be remote areas, typically 2 to 5 km away from any major roads and at least 500 m upwind of even minor roads or pathways. Generally, one location was sampled during each day and the distance between consecutive sites was about 150 km.

[8] The altitude of the sampling location was ∼<10 m along the coast, varied between 200 to 600 m in the inland region; ∼1 km in the Western Ghats and the highest sampling altitude was Kodaikanal (KKN, 10.22°N, 77.4°E) at 1880 m above msl. As the Aethalometer was sampling at a constant mass flow rate, the estimated BC values were corrected for the change in the pump-speed arising due to changes in the ambient pressure from location to location (due to change in altitude) following the details given by Moorthy et al. [2004].

3. Results and Discussions

3.1. Spatial Variation of BC Mass Concentration

[9] By averaging the individual MB measurements made during the period 09:00 to 16:00 hrs (when the diurnal variations are either absent or insignificant) on each day, daytime mean MB is estimated, which generally corresponded to the lowest concentration of BC during the day. Considering these as spatial samples obtained from a temporally stable (statistically) population, a spatial composite is generated as shown in Figure 2. Generally, high BC concentrations (>2.5 μg m−3) are observed due south of ∼14°N, with very high values (>4 μg m−3) along the coastal region (Mangalore, MNG) and the southern hilly, forest regions of the Western Ghats. These values are comparable to those seen at inland urban centre like Hyderabad; even though the coastal regions are much less urbanised (compared to centres like Hyderabad and Bangalore). Thus, the higher density of population, industrialisation (fertilizer, oil refineries, and thermal power plants) and the presence of three major harbours (Karwar, Mangalore and Kochi) along the coast might be contributing to the high BC values along the coastal regions. Also the prevailing low-level winds show a convergence over the western coastline, changing over from near easterlies to near northerlies, which could favour spatial confinement of aerosols [Moorthy et al., 2005].

Figure 2.

The spatial variation of BC mass concentration (MB).

[10] The very high BC, between 9 to 11°N, 77°E, in the Western Ghat region where the human activities are sparse, is associated with the extensive forest fires that occurred during February–March 2004 in which around 27000 hectares of forest and grass land were affected (reports of the forest department). Some of these fire-affected regions were spotted by Team-1 during their return.

[11] Compared to the above, the regions due north of 14°N, show only moderate levels of BC (between 1 to 2 μg m−3) up to 18°N except at the urban centre Hyderabad (17.3°N, 78.5°E) where MB > 4 μg m−3. This high BC is associated with the urban activities and high density of transport in the city of Hyderabad, which is the fifth largest city in India [Lata et al., 2004]. To the north of 18°N, BC concentrations are quite low (<1 μg m−3) and are typical to values observed over remote continents/polluted oceans [Andreae et al., 1984].

[12] The atmospheric distributions of BC are largely determined by circulation and the atmospheric abundance of BC is primarily influenced by precipitation; both were quite weak over the study area, during the study period. As such, concentration of BC over India is likely to be high during the winter season. Babu and Moorthy [2002] reported that during the dry months (December to March) the mean BC concentration is ∼5 μg m−3 at Trivandrum, a coastal station devoid of major industries and attributed it to the combined effects of the longer residence time and advection by the prevailing winds. Based on INDOEX measurements from Goa, lying upwind of Mangalore, Leon et al. [2001] reported an average value of 3 ± 0.7 μg m−3 for the winter months. Babu et al. [2002] had reported an average value of 4 μg m−3 for Bangalore, which is an urban station, in the peninsula. Values in the present investigations are, in general, in-line with these when similar environments are compared.

[13] As the spatial sampling was confined to a narrow longitudinal belt of ∼4° width, but spread over a long latitudinal range from 8°N to 22°N (Figure 1), the data were used to examine the latitudinal variation of MB, by grouping the individual measurements into latitudinal bins of 1° (irrespective of longitude) and averaging. The resulting pattern is shown in Figure 3. Despite the scatter, (mainly due to the forest fire infested regions down south) Figure 3 indicates a decrease in MB from south to north, the regression slope yielding a decrease of ∼160 ng m−3 for every degree increase in the latitude, with a correlation coefficient of 0.79. The fairly large scatter at the lower latitudes (below∼12°) is also caused by the large variation of the environmental conditions (from the densely populated and industrialized coastal land to sparsely inhabited and vegetated forest land to the east), which is a characteristic feature of this region. A word of caution, however, is warranted here. Our study is limited to Peninsular India and does not include the Indo-Gangetic Plain (IGP). The northern end of our track just touches the southern edge of the IGP. Several recent measurements have shown very high concentrations of aerosols and BC over the IGP during winter [e.g., Tripathi et al., 2005] with BC concentrations going up to 20 to 30 μg m−3 and contributing ∼7–15 % to the total aerosol mass.

Figure 3.

Latitudinal variation of BC mass concentration (MB). The points are the means and the vertical bars through them are the standard deviations of the means.

3.2. Spatial Variation of BC Mass Fraction

[14] From the collocated measurements of MB and MT it was possible to examine the spatial variation of BC mass fraction FBC(= MB/MT); which was estimated for each location, as an average for the daytime period between 09:00 and 16:00 LT. The spatial variations of FBC, shown in Figure 4, differ from that of MB in Figure 2, indicating that the FBC need not necessarily follow MB. Even though FBC tends to be high (∼8 to 14%) at regions of high MB, the high FBC regions are comparatively more spread spatially implying higher fractional load even at locations where the absolute concentration of BC is lower. This might be caused by the finer size of BC aerosols and the resulting longer lifetime of BC (compared to composite aerosols), which favours wider spatial dispersion of the species, by the prevailing winds, to locations farther away from the potential sources. Another feature seen is the extremely high FBC over the Ghat regions near Kodaikanal, where the low MT (primarily due to the elevation, forested and sparsely inhibited nature) and high MB (associated with this forest fires) lead to very high FBC (∼30%). In contrast, very low BC mass fractions (<4%) are noticed around the area 15 to 17.7°N and 74 to 77°E and also at 22°N. Nevertheless, there are regions within the central peninsula, which are not apparently highly industrialised or urbanised, where BC constitute to >10% of the ambient aerosols. This high BC is of concern from climate as well as environmental and health perspectives. The high BC mass fraction observed in the central interior regions could be due to (1) transported BC owing to its lifetime of about 6 to 7 days, (2) reduced wet removal, (3) local generation of BC, and (4) the lower values of composite aerosol loading due to reduced human activities.

Figure 4.

The spatial variation of BC mass fraction (FBC). The points correspond to the measurement locations, and the height of the bar on each point is proportional to the FBC measured at that location.

[15] The contribution of BC in aerosol reported during ACE-2, INDOEX and TARFOX were in the range of 10 to 15% [Novakov et al., 2000; Ramanathan et al., 2001] where as that reported over tropical western Pacific was about 25% [Liley et al., 2002]. During the field campaign at an urban station Bangalore (13°N, 77°E, 960 m msl), large amount of BC was observed; both in absolute terms and fraction to total aerosol mass (∼11%) and fine aerosol mass (∼23%) [Babu et al., 2002]. Based on the measurements from Trivandrum, Babu and Moorthy [2002] reported large seasonal variation in FBC from a high value (11%) in winter to a low value (3 to 4%) in monsoon season. Satheesh et al. [1999] observed BC mass fraction of 6% over the Indian Ocean during INDOEX.

[16] The latitudinal variation of FBC, shown in Figure 5 (obtained similar to MB), also differs significantly from that of MB. Unlike the case with MB, which showed a rather monotonic decrease towards north, FBC shows an oscillatory variation with an overall decrease up to ∼16°N, where it reaches to ∼4%. A few ups and downs follow before it eventually reached the lowest value (∼3%) beyond 20°N. Even though the highest values of FBC at ∼10°N is ignored as due to the rather localised forest fire, high FBC still persisted in the southern peninsula. The rate at which FBC decreases towards north is slower than that of the MB. Despite, a spatially averaged value (excluding Kodaikanal) works out to be ∼6.1%, which is rather high; both from the perspectives of radiative forcing and air quality. The slower variation, with latitude, of FBC (compared to MB) is attributed to the different generation mechanisms of the composite and BC aerosols. Moreover, the atmospheric residence time of BC is much longer than that of the composite aerosols, where the coarse mode contributes significantly due to the dry season and semi-arid terrain.

Figure 5.

The latitudinal variation of BC mass fraction (FBC).

3.3. Implications to Radiative Forcing

[17] The high values of FBC observed in our study have important implications to radiative forcing, over land. Based on model calculations over ocean, Babu et al. [2004] have shown that the AOD remaining the same, increase in FBC leads to an increase in the short wave radiative forcing of the atmosphere. In the present study the value of FBC varied between ∼3 and 8%. These extreme values are respectively characteristic to the ‘continental average’ and ‘urban aerosol’ models of Hess et al. [1998]. The terrain varied from vegetated forest to semi-arid. Keeping these extreme cases, we have estimated the range of short wave radiative forcing efficiency for the above two models. The results show that the TOA forcing efficiency varies between ∼−10.2 W m−2 (3% BC, vegetated terrain) and ∼+10.9 W m−2 (8% BC urban locations). Corresponding atmospheric forcing efficiencies are +52.4 W m−2 and +103.9 W m−2. This large change in the forcing efficiency due to the spatial variation in FBC shows the inadequacy of single point measurement for regional impact assessment. More accurate assessments should consider the vertical distribution of BC also.

4. Summary

[18] (1) High BC mass concentrations (>2.5 μg m−3) are seen along the west coast, moderate values (1 to 2.5 μg m−3) from 15 to 18°N, and very low values (<1 μg m−3) due north of 19°N.

[19] (2) Besides being high (>8% of the composite aerosol mass) along the regions of high MB, the high FBC is more spread to the interior plateau regions.

Acknowledgments

[20] This work was carried out during the land campaign under Indian Space Research Organization's Geosphere Biosphere Program. We thank Preetha S. Pillai, Auromeet Saha, and K. Madhavi Latha for their help during the data collection and analysis.