Temporal variability of MODIS aerosol optical depth ... - Springer Link

Environ Sci Pollut Res (2015) 22:1329–1343 DOI 10.1007/s11356-014-3418-2

RESEARCH ARTICLE

Temporal variability of MODIS aerosol optical depth and chemical characterization of airborne particulates in Varanasi, India Vishnu Murari & Manish Kumar & S. C. Barman & T. Banerjee

Received: 10 May 2014 / Accepted: 6 August 2014 / Published online: 21 August 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Temporal variation of airborne particulate mass concentration was measured in terms of toxic organics, metals and water-soluble ionic components to identify compositional variation of particulates in Varanasi. Information-related fine particulate mass loading and its compositional variation in middle Indo-Gangetic plain were unique and pioneering as no such scientific literature was available. One-year ground monitoring data was further compared to Moderate Resolution Imaging Spectroradiometer (MODIS) Level 3 retrieved aerosol optical depth (AOD) to identify trends in seasonal variation. Observed AOD exhibits spatiotemporal heterogeneity during the entire monitoring period reflecting monsoonal low and summer and winter high. Ground-level particulate mass loading was measured, and annual mean concentration of PM2.5 (100.0±29.6 μg/m3) and PM10 (176.1±85.0 μg/m3) was found to exceed the annual permissible limit (PM10: 80 %; PM2.5: 84 %) and pose a risk of developing cardiovascular and respiratory diseases. Average PM2.5/PM10 ratio of 0.59±0.18 also indicates contribution of finer particulates to major variability of PM10. Particulate sample was further processed for trace metals, viz. Ca, Fe, Zn, Cu, Pb, Co, Mn, Ni, Cr, Na, K and Cd. Metals originated mostly from soil/earth crust, road dust and re-suspended dust, viz. Ca, Fe, Na and Mg were found to constitute major fractions of particulates

Responsible editor: Constantini Samara V. Murari : M. Kumar : T. Banerjee (*) Institute of Environment and Sustainable Development, Banaras Hindu University, Varanasi 221005, India e-mail: [email protected] T. Banerjee e-mail: [email protected] S. C. Barman Environmental Monitoring Division, Indian Institute of Toxicology Research, Lucknow, India

(PM2.5: 4.6 %; PM10: 9.7 %). Water-soluble ionic constituents accounted for approximately 27 % (PM10: 26.9 %; PM2.5: 27.5 %) of the particulate mass loading, while sulphate (8.0– 9.5 %) was found as most dominant species followed by ammonium (6.0–8.2 %) and nitrate (5.5–7.0 %). The concentration of toxic organics representing both aliphatic and aromatic organics was determined by organic solvent extraction process. Annual mean toxic organic concentration was found to be 27.5±12.3 μg/m3 (n=104) which constitutes significant proportion of (PM2.5, 17–19 %; PM10, 11–20 %) particulate mass loading with certain exceptions up to 50 %. Conclusively, compositional variation of both PM2.5 and PM10 was compared to understand association of specific sources with different fractions of particulates. Keywords Particulate . Aerosol optical depth . Metal . Organics . Water-soluble ions . Varanasi

Introduction In recent decades, anthropogenic activities have altered the intrinsic properties of atmosphere either through trace gases and particulate emission or by reducing their global sinks. This perturbation is of such magnitude that it is expected to lead substantial implications for the future climate (Banerjee et al. 2011a). Among several identified species, methane, oxides of nitrogen and carbon, ozone and airborne particulates are typically reactive and provide substantial radiative impacts on regional to global climate. However, until 1980s, there was little unanimity regarding climatic importance of airborne particulates as their influence was far more difficult to assess than that of greenhouse gases. Subsequently in late 1980s, it has been understood that airborne particulates are significant in mediating physical and biogeochemical exchanges between the atmosphere, land surface and ocean. Airborne particulates

1330

directly influence climate either by scattering and absorbing incoming solar radiation or indirectly through modifying cloud micro-physical processes (Ramanathan and Feng 2009). Several field campaigns and collaborative research programmes initiated from 1990s subsequently revealed that most significant factor that regulates the climatic potential of particulate is not its mass loading but its chemical nature. Airborne particulates are multi-component mixtures which originate from a range of sources (direct emission or through gas to particle conversion) and contribute adversely to health, crop yield, visibility and many other sectors of the climate system. The fine mode particulates (PM2.5) are mainly originated through gas to particle conversion mechanism (fossil fuel, biomass combustion and anthropogenic emission), while coarser particulates (PM2.5–10) arise from natural sources (Aldabe et al. 2011; Banerjee et al. 2011a, b). However, nature and physical behaviour of particulates includes heterogeneity at spatial levels and hence their effects vary with topography, climate and meteorological conditions. The chemical nature of particulates at the middle Indo-Gangetic (IGP) region is mostly characterized by the presence of organic aerosols generated through burning of biomass and fossil fuel, which further gets complicated through mixing of trans-continental haze coming from the Thar deserts and surrounding area. Thus, in middle IGP, information related to particulate mass concentration, variability in their geographic distribution and nature of interactions still remain a complication (Ramachandran et al. 2012). Such creates the essentiality of scientific understanding of particulate’s sources, characteristics and spatiotemporal distributions at middle IGP which will eventually help to identify the possible vulnerable sectors which need to be prioritized for climatic adaptation. Additionally, particulateinduced health hazards have also raised the essentiality of conducting elaborative research relating tropospheric particulate and its chemical nature (Viana et al. 2008). Morphology and chemical compositions of atmospheric particles play the decisive role in its transport, transformation and removal mechanism (Sharma and Maloo 2005). Therefore, in recent years, there has been much interest in chemical fractionation analysis of airborne particulates, which also helps to identify associated sources and eventually assist source-specific adaptation and management policies (Continia et al. 2014). The bulk component of atmospheric particulate load consists of ammonium salts, nitrate, sulphate, elemental carbon and organic material, the latter being made up of different individual organic compounds, viz. alkanes, alkenes, fatty acids, ketones, aromatics, alcohols, aliphatics and sugars (Ciaccio et al. 1974; Sharma and Maloo 2005). Organic matter shares the predominant section of atmospheric particulate load (PM10: 10–40; PM2.5: 30–60 %) and held responsible for its mutagenicity (Crebelli et al. 1991), therefore, needs to be measured and characterized. For the present research work, sampling and gravimetric determination of benzene-soluble

Environ Sci Pollut Res (2015) 22:1329–1343

airborne organic compound associated with particulate matter was measured as an indicator of toxic organic fractions which directly relates it with associated health hazards (Ciaccio et al. 1974; Sharma and Maloo 2005; ASTM 2010). It is noteworthy to mention that ambient level of airborne particulates and its chemical composition with some degree of certainty has extensively done for the most part of urban India (Pant and Harrison 2012), but there are very few instances where particulate concentration are reported for middle IGP, especially in the context of Varanasi. Particulate concentrations reported earlier mostly exceeded the national permissible limit (Fig. 1) which pose a serious threat to local habitat. However, till the submission of the manuscript, there was no such information available regarding ambient concentrations of finer particulates (PM2.5) and its compositions for the particular region. In this regard, this manuscript will be first of its kind to report the aerosol chemical (metals and ions), optical characters (AOD) with associated organics (BSOF) in an urban area of middle IGP. The field campaign carried out from March, 2013 to February, 2014 will possibly explain the concentration variation of airborne particulates, associated toxic organics and compositional variation of metals and water-soluble inorganic ionic components (WSIC) both in terms of PM2.5 and PM10.

Methodology Study area One-year air quality assessment in terms of airborne particulates was carried out from March, 2013 to February, 2014 at Varanasi, the spiritual capital of India, situated in the middle Gangetic plain of the Indian sub-continent (Fig. 1). Varanasi with an estimated area of 225 km2 is located at 25° 18′ N latitude, 83° 01′ E longitude and 82.20 m above sea level with a population of 3.4 million. Although the city itself does not possess any significant industrial activities, but road dust re-suspension, extensive commercial activities and vehicular exhausts may be considered as the dominant sources of regional pollutants. Additionally, a large number of small scale industries, viz. food processing; fabric printing and dyeing; paint manufacturing; batteries; electrical cables; newspaper printing and various manufacturing industries including iron gates, window nets, bicycle tires and heavy agricultural equipment are dispersed in west and north west of the city. Keeping in view the environmental significance of the potential influence of toxic metals and ions associated with particulate, and their continued high levels in the local atmosphere, the principal objective was to evaluate the atmospheric abundance of trace metals and ions in airborne particulates at Varanasi.


1331

Fig. 1 Spatial distribution of particulate concentrations in and around monitoring location. a Pandeypur (Prajapati and Tripathi 2008; Pandey et al. 1992); b Rajghat (Pandey et al. 1992); c Cantt (Prajapati and Tripathi 2008); d Lahartara (Pandey et al. 1992); e Godolia (Pandey

et al. 1992); f Lanka (Prajapati and Tripathi 2008); g BHU campus (Pandey et al. 1992) PM10 values are underlined, rest are total suspended mass. All concentrations are expressed in μg/m3

Micro-meteorology

maximum of 7.4 km h−1 for the entire monitoring period. The annual variation of wind vector has been illustrated in Fig. 2. Wind direction was predominant from westerly during summer and winter seasons whereas easterly during monsoon.

The climate of Varanasi is characterized as humid sub-tropical, with high temperatures (38.5 to 41.2 °C) during summer (March to June), intense rainfall during monsoons (July to September, annual rainfall of 1,100 mm, 90 % of which occurs in monsoon) and severe cold during winter seasons (December to February, 8.4 to 15.0 °C). The daily sunshine period varied from 7.2 to 10.7 h during summer and 6.4 to 8.7 h during winter. The average maximum relative humidity varied between 11.1 and 92.2 % during summer and 57.0 to 100.0 % during winter. Wind speed varied from a minimum of 1.5 to a

Trends in aerosol optical depth over India Seasonal and annual trends in aerosol optical depths (AODs) during the entire ground-level particulate monitoring period (February, 2013 to March, 2014) were retrieved using Moderate Resolution Imaging Spectroradiometer (MODIS) Level 3 (1×1°) remote sensing data. The MODIS operating at an

1332


Fig. 2 The annual variation of wind vector for the entire monitoring period at BHU, Varanasi

altitude of 705 km measures reflected solar radiance and terrestrial emission (resolutions between 250 m and 1 km) which directly relates it with columnar aerosol loading with nearly global coverage at moderate spatial resolutions (Ramachandran et al. 2012). Therefore, particle information retrieved by satellite sensors may be suitable for assessing spatial and temporal trends of finer particulates over large geographical areas (Liu et al. 2007). For the present experiment, MODIS remote sensing information on board the Earth Observing System (EOS) Terra satellites was solely used. Satellite-based Level 3 MODIS data were obtained from MODIS online visualization and analysis system (http://disc. sci.gsfc.nasa.gov/giovanni) by averaging daily AOD at 550 nm over the Indian sub-continent (collected from March, 2013 to February, 2014) and presented in Fig. 3. Additionally, aerosol extinction or AOD value has also expressed in terms of three prominent seasons (summer, monsoon and winter) to identify seasonal variability, transport and transformation of airborne particulates.

used for PM10 sampling. APM-460BL was operated at a constant flow rate of 1 m3 h−1 on EPM-2000 filter paper (Whatman) of 20.3×25.4 cm (8″×10″) size. Initially, 24h desiccated filter papers were weighed twice on the balance (AY220, Shimadzu). Conditioned and weighed filter papers were placed in a filter holder (for PM2.5) and cloth-lined envelope (for PM10) and taken to the field for sampling to avoid any possible contamination. Both samplers were placed at the height of 7.5 m in the campus of Banaras Hindu University on the roof of the Institute of Environment and Sustainable Development. On completion of sampling, filter papers were removed from the filter holder by non-serrated plastic tweezers and transferred into desiccator for 24 h. After final weighing, the filter paper was placed into cassette and wrapped in aluminium foil to prevent exposure to sunlight and photo-oxidation of samples. Exposed filters were stored into the fridge at 4 °C until chemical speciation was done.

Particulate collection and analysis Chemical speciation of particulates Ambient air quality monitoring for airborne particulates was carried out twice a week continuously for 24 h (10:00 – 10:00 h). Sampling of fine particulates (PM2.5) was done by fine particle sampler (APM 550, Envirotech) in EPM-2000 filter paper (Whatman, 47-mm diameter) with a continuous flow rate of 1 m3 h−1. High-volume particle sampler (APM460 BL, Envirotech) with size-selective inlet was additionally

The climatic effects of particulate are highly attributed to its sizes and morphology due to a variety of compounds they contain and associated sources. In order to establish sourcespecific relations and spatiotemporal distributions of particulates, chemical characterization is essential. Additionally, chemical characterization also helps to establish specific


1333

Fig. 3 Aerosol optical depth (AOD) retrieved from MODIS in Indian sub-continent (a) and its seasonal variations (b–d)

relations between particulates and its associated health impacts (Banerjee et al. 2011a, b; Sharma and Maloo 2005). Toxic organic content To identify the toxic organic fractions associated with particulates, ASTM (ASTM 2010) guidelines were used. It refers to gravimetric estimation of the benzene-soluble particulate matter whose carcinogenicity has already been established by several scientists (Ciaccio et al. 1974; Crebelli et al. 1991). The benzene extracts can either be analysed on various platforms for speciation or subjected to gravimetric analysis for bulk measurements. BSOF measurements were carried out for both kinds of particulates; however, this particular manuscript only reports the organics associated with PM 10 expressed as benzenesoluble organic fractions (BSOF). Detailed chemical

characterization of BSOF for both PM10 and PM2.5 is under process through gas-chromatography mass spectroscopy analysis. One-fourth (2″×2.5″) of exposed EPM-2000 filter was cut into smaller parts in a cleaned, dried beaker, and 30 mL ultrapure (HPLC grade) benzene was added. Beaker was sealed immediately by aluminium foil to minimize loss of organics while ultrasonication. It was ultrasonicated for 5 min at room temperature and kept ideal for 30 min for complete dissolution of organic compounds. The above process has been repeated before eventually being vacuum filtered to a clean, oven-dried pre-weighted flask. The solvent was evaporated to dryness in a pre-heated oven at 40 °C for 15–20 h. The beaker was re-weighed (AY 220, Shimadzu) after cooling, and difference between final and initial weight represents the organic fractions associated with particulates. The air

1334


concentration of soluble particulate organic fraction was calculated as ðA−BÞ 5 ⋅ Concentration of BSOF μg=m3 ¼ V F

Where, A B V F

is the residue weight of sample, mg residue weight of blank, mg volume of air sample, m3 volume of aliquot, mL

NaOH (50 % w/w) with triple-distilled water as regenerator. Cations, viz. NH4+, K+, Ca2+ and Mg2+ were detected using a suppressor (CSRS-300, 4 mm; Dionex, USA) with a separation column (IonPac CS17-HC, 4×250 mm; Dionex, USA) with a guard column (IonPac CG17-HC, 4×50 mm; Dionex, USA) and 5 mM MSA (methane sulphonic acid) as eluent. The IC system was fitted with a 25-mL sample loop that was used to introduce the sample manually. Chromatography data were collected at 5 Hz, and chromatograms were processed using the Chromeleon software. Several blank filters were also analysed for cations and anions, and analytical error (repeatability) was estimated to be 3 % based on triplicate analysis.

Trace metals The extraction and analysis of metals was carried out as per the US EPA Method IO-3.2 (EPA 1999). As per reference, filter paper handled with the use of non-serrated plastic forceps, one-fourth of filter paper was cut into pieces in a beaker and was place in a strip-down position ensuring the entire exposed portion was covered by extracting solution (5.55 % HNO3 with 16.67 % HCl). Extracting solution containing PM10 and PM2.5 refluxed gently on the hot plate by covering it with watch glass until the solution becomes clear. Digested sample was filtered (No. 42, Whatman), made up to require by Milli-Q water and stored in freezer for further analysis. The filtrate was examined for the concentration of Ca, Fe, Zn, Cu, Pb, Co, Mn, Ni, Cr, Na, K and Cd by atomic absorption spectrophotometer (Avanta Ver 2.01, GBC). Instrument was calibrated thrice for each metal using known certified reference material (Qualigens make) before analysis. In order to examine the background heavy metal content, identical extraction and analysis procedure was employed for blank filter papers. Water-soluble ionic constituents (WSIC) The water-soluble components associated with both PM10 and PM2.5 were extracted through de-ionized Milli-Q water having a conductivity of >18.2 MΩ for 90 min (35 °C) in 25-mL polypropylene tubes by ultrasonication. Ultrasonic method is accepted for the extraction of inorganic ionic components, and normally over 98 % of sulphate, nitrate, ammonium can be extracted through this process (Willeke and Baron 1993). The ultrasonicated sample was further subjected to filtration (0.45-μm pore size, Millipore), and WSIC content of PM2.5 and PM10 has been analysed by ion chromatograph (ICS3000, Dionex, USA). Initially, concentrations of F−, Cl−, NO3− and SO42− were determined by ion chromatograph (IC) using an anion micro-membrane suppressor (ASRS300, 4 mm; Dionex, USA) with IonPac AS11-HC×250-mm analytical column. The eluent for anion analysis was 20 mM

Results and discussion Aerosol optical depth (AOD) In order to assess the particulate loading at IGP for the entire monitoring period, Moderate Resolution Imaging Spectroradiometer (MODIS) data on board Terra satellite were analysed for aerosol optical depth at 550 nm. The AOD directly relates the optical properties of atmospheric aerosol in a column-integrated manner (top of atmosphere to land surface) and mainly influenced by type of particulates present instead of mass loading (Kaufman et al. 2002). Therefore, ground-level measurement of fine and coarse particulates may not be always well synchronized with satellite AOD. However, there are few instances which report the presence of linear relationships between MODIS AOD and groundlevel finer particulates (Liu et al. 2007; Wang and Christopher 2003). For the present study, representation of seasonal variability and trans-boundary movement of airborne particulates at IGP was only intended. The general aerosol characteristics over the IGP for the entire monitoring period are shown in Fig. 2a with an indication of ground-level particulate monitoring station. Figure clearly indicates that AOD over IGP is much higher than the rest of India. Higher AOD may be the consequence of desert dust influx originated from the western arid regions of Africa, Middle-East, and Thar (Rajasthan) regions, predominately during summer and pre-monsoon season (El-Askary et al. 2006). Additionally, recent urbanization and industrial development at IGP, regional practices of biomass burning coupled with emission from residential biofuel (wood, crop residue, dung) are the largest source of atmospheric aerosol concentrations. Spatial heterogeneity in AOD was also observed over the IGP both during summer (Fig. 3b) and winter (Fig. 3d). A plausible explanation of such observation may be frequent trans-boundary movement of dust storms originated over Thar


Desert. Additionally, there might be implications of atmospheric boundary layer (ABL) variations which strongly restricts atmospheric ventilation (Banerjee et al. 2011a, b). During summer (Fig. 3b), the ABL height increases with strong insolation coupled with hygroscopic growth of water-soluble aerosols result in higher AOD (Ramachandran et al. 2012). Contrastingly, ABL remains low during winter which results in the formation of inversion layer which restricts particulate vertical mixing (Banerjee et al. 2011a, b). The shallow boundary layer and less ventilation coefficient facilitate the trapping of particulates resulting in higher AOD. Figure 3c represents lower AOD level during monsoon period which clearly indicates particulate wet scavenging through monsoonal rain. The summer monsoon season (July to September) accounts for 80 % of the total annual rainfall in India which may be well responsible for the removal of mass particulates from the lower atmosphere resulting in lower AOD. However, according to Ramachandran et al. (2012), correlation hardly exists between trends in AOD and monsoonal rainfall due to the differences in the spatiotemporal variations in rainfall and the particle diversity. Temporal variation of particulates Annual variation of ground-level PM10 concentrations is shown in Fig. 4 for the entire sampling period. Annual mean concentration of PM10 was 176.1±85.0 μg/m3 (mean ± sd) (n =104), whereas monthly average concentration varies within 43.6–318.5 μg/m3. Diurnal variation of particulates was 12 to 409 μg/m3, reflecting significant variability. It is evident that annual average PM10 not only exceeds the annual permissible limit (NAAQS, 60; WHO, 20 μg/m3), often it persists (80 % over NAAQS and 100 % over WHO standard) in a range of 8– 9 times over than the prescribed standard, which may significantly contribute to the risk of developing cardiovascular and respiratory diseases as well as of lung cancer. Although air

1335

pollution-health interaction information is extremely rare in Indian context, however, according to WHO report (WHO 2013), outdoor air pollution caused 0.62 million premature deaths in India in 2010, a staggering increase of sixfold than that of 2000. As evident from AOD, ground-level PM10 concentration follow the identical dynamic variability during the entire monitoring period reflecting minimum mass loading during July to September, 2013 due to wet scavenging of particulates by monsoonal rainfall. Contrastingly, summer (March to June, 2013) characteristically exhibits higher ground-level PM10 mass loading triggered by generation and persistence of Asian aerosol from Middle-East countries and Thar Desert, Rajasthan, India, which is further aggravated by the development of hygroscopic growth of water-soluble aerosols. Finer particulate (PM2.5) contributes dominantly to visibility degradation and incurs changes in regional radiative balance and are of special interest when health problems are concerned. Figure 5 presents the monthly average PM2.5 levels at the sampling location which varies between 50.1 and 154.0 μg/m3 with an annual mean of 100.0±29.6 μg/m3 (n=104) persisting well above than the annual permissible limit (NAAQS, 40 μg/m3; WHO, 10 μg/m3). Annual average PM2.5 concentration persists only 16 % below the NAAQS, whereas concentrations remain always higher than the WHO standards. Additionally, PM2.5 concentration variations resemble the same pattern exhibited by PM10 and satellitemeasured AOD. This signifies that both types of particulates originate from and are regulated by the same type of external sources. Finer particulates (PM2.5) contribute a fraction of coarser suspended particulates (PM10), and therefore, it is expected that both particulates demonstrate a close statistical relation. There are references available that demonstrate close ratio between PM10 and PM2.5 concentrations (Keywood et al. 1999). Additionally, average and standard deviation (Sd) of

Fig. 4 Temporal variation of PM10 concentration at ground monitoring station

1336


Fig. 5 Temporal variation of airborne finer particulate (PM2.5) concentration at ground monitoring station

PM2.5/PM10 ratio are also considered to be good indicators of the relative contributions of finer particulates to PM10. For the present study, daily variations of PM2.5/PM10 persist 0.35 to 0.90, with an average of 0.59±0.18 (Fig. 6). This indicates that the proportion of fine particulate to PM10 has seasonal variability, and for most of the monitoring periods, finer particulates contribute major variability to PM10. Higher ratios (>0.5) are mostly observed during summer and post-monsoon season in contrast to winter months where lower ratios were observed. Figure 7 shows a linear bivariate plot of PM2.5 versus PM10 with computed coefficient of determination (R2 =0.481) with correlation coefficient (r) of 0.71. It clearly indicates that a certain degree of correlation exists between finer and coarser particulates, which may be specifically due to common source which influence origin of both kinds of particulates. Temporal variation of toxic organic compounds The concentration of organics associated with particulates is usually determined by organic solvent extraction of samples

Fig. 6 Temporal variation of PM2.5/PM10 ratio at ground monitoring station

collected on glass fibre filters (ASTM 2010). This organic fraction of the particulate matter represents both aliphatic and aromatic organics in a given weight of particulates. Although there are no statutory limits of ambient BSOF concentrations, workplace standard for BSOF has been reported as 200 μg/m3 by Mastrangelo et al. (1996) and well accepted by Sharma and Maloo (2005). However, realistical ambient standard should be much less than that of industrial standard, probably onetenth, which leads to a value of 20 μg/m3 as a safe value of ambient BSOF concentrations. The monthly minimum, maximum and average concentrations of BSOF associated with PM10 are shown in Fig. 8, and relative comparison of annual variation of PM2.5 and PM10 with associated toxic fractions are presented in Fig. 9. Annual mean BSOF concentration was 27.5±12.3 μg/m3 (n=104) with monthly average concentration varying within 8.2– 48.6 μg/m3. Daily average concentrations of BSOF was also found to exceed 62 % time above the acceptable level (20.0 μg/m3) indicating the presence of high levels of organic compounds including PAHs. Detailed characterization of toxic organic fractions associated with particulates is under


1337

Fig. 7 Scatter plot between PM2.5 and PM10 at ground monitoring station

process through chromatographic techniques. For the entire sampling period, BSOF constitutes 11–20 % of PM10 mass loading with certain exceptions up to 50 %. However, in contrast to other metros in India, viz. Delhi (BSOF, 48 μg/ m3, unpublished data), Kanpur (BSOF, 106.71±62.38 μg/m3, Sharma and Maloo 2005) and Kolkata (BSOF, 45.63 ± 27.68 μg/m3, Gupta et al. 2006), toxic organic fractions was lower in Varanasi. Results were inherently consistent with expectations as Kanpur is one of India’s highly polluted city, and heavy traffic and adjacent industrial activities may be the reasons behind elevated toxic organics in Kolkata and Delhi. Trace metals Characterization of trace metals composition concentrated in ambient particulates is presented in respect to their associated

sources. Trace metals was found to be present in significant amount in both coarser and finer fractions of particulates (Fig. 10a, b). Metals that originated mostly from soil/earth crust, road dust and re-suspended dust, viz. Ca, Fe, Na and Mg (Fujiwara et al. 2011; Gu et al. 2011) were found to be much high in concentrations in coarser fractions; however, their presence in finer particulate also found to be significant. Airborne calcium concentrations were in between 7.01 and 46.00 μg/m3 (PM10) and 6.10 and 35.00 μg/m3 (PM2.5), contributing a significant proportion (4–5 %) of particulate mass loading. Contrastingly, sodium and iron concentrations were mostly prevalent in PM10 (Na: 0.11 to 43.98 μg/m3; Fe: 0.62 to 14.22 μg/m3) rather than PM2.5 (Na: 0.06 to 9.33; Fe: 0.43 to 7.20 μg/m3). Such was according to expectations as these metals are universally accepted as crustal markers and, therefore, mostly associates with coarser particulates. Although Na has the major marine

Fig. 8 Temporal variation of toxic organic fractions (BSOF) associated with PM10

1338

Fig. 9 Relative comparison of variation of PM10, PM2.5 and toxic organic fractions for the entire monitoring period

source, but it has been considered as crustal markers as interference of marine aerosol in Varanasi is negligible. Potassium (K) is used as a marker of crustal dust in coarser particulates while soluble K+ for biomass burning in the fine range of particulates. In India, K has been used an elemental marker for biomass/wood combustion (Khare and Baruah 2010), whereas zinc (Zn) is considered to originate from refuse burning and incineration of hazardous waste (Bullock et al. 2008). Biomass, refuse and waste burning are common practices of the inhabitants of the IGP region; therefore, high concentrations of both K and Zn levels were expected in Varanasi. For both K (PM2.5: 0.18 to 10.12; PM10: 0.20 to 12.73 μg/m3) and Zn (PM2.5: 0.11 to 3.20; PM10: 0.22 to 14.87 μg/m3), experimental results were in accordance to expectations; however, characteristically both metals were found to be mostly associated with coarser particulates. Metals arising from automobile sector are normally contributed by tailpipe emissions, wear and tear of tires and brakes and re-suspension of road dust. According to international reviews, elemental markers for such emission include Cu, Pb, Ni and Mn (Begum et al. 2011). However, since the introduction of unleaded petrol in India from 1990s, atmospheric Pb is disregarded as unique marker for vehicle emissions. Atmospheric abundance of Pb in Varanasi was measured below national standard (500 ng/m3), but characteristically both types of airborne particulates (PM2.5: 12.3 to 137.2 μg/m3) (PM10: 16.4 to 164.4 μg/m3) were found to be associated with Pb. This clearly indicates that Pb has originated from common sources which directly influence both PM2.5 and PM10. Concentrations of Ni mostly persist within below detection level (BDL) of 20 ng/m3 for both kinds of particulates with occasional increase above annual standards (20 ng/m3). However, finer particulates were predominately found to be associated with Ni. Anticipated sources of Ni in ambient air of Varanasi may be vehicular


emissions; refuse burning; incineration; hazardous waste disposal and industrial emissions, viz. fabric printing and dyeing, paint manufacturing and batteries. The most specific metals that are accepted as a marker of industrial emissions are Co (Kar et al. 2010), Cr (Shridhar et al. 2010) and Cd (Mouli et al. 2006). Co and Cr were present in a range of BDL to 50 ng/m3. Additionally, variability of concentration mostly persists within a narrow range of BDL to 27.6 ng/m3, which clearly indicates that such trace metals were less abundant in ambient air. However, Cd was more frequent than the remaining two trace metals and was found to present within BDL to 51.4 ng/m3 with higher variability. As expected, all these trace metals were mostly associated with finer particulates establishing their anthropogenic origin. It is expected that paint-battery industry and automobile emissions are the possible sources of Cr and Cd found associated with airborne particulates. Water-soluble ionic constituents (WSIC) Water-soluble ionic constituent (WSIC) found in different size fractions of airborne particulates are reported in Fig. 11. However, it is noteworthy to mention that although Ca was processed both by IC (as WSIC) and AAS (metals), for the present manuscript, only total Ca is reported. The analysed ions accounted for approximately 25 % (PM10: 26.9; PM2.5: 27.5 %) of the particulate mass concentration, with sulphate as the dominant species since on average it contributed almost 9 % to the measured concentration. For both fractions of particulates, SO4−2, NH4+ and NO3− formed the major components of WSIC followed by other ions, i.e. K+, Mg+2, F− and Cl−. In PM10, SO4−2 was reported to have a maximum concentration of 17.62 (±1.11)μg m−3. Sulphate in the atmosphere is mainly attributed to gas to particle conversion initiating from SO2 coming out of different sources including industrial, power generation as well as domestic activities. Nitrate forms the second largest fraction of WSIC with an atmospheric concentration of 15.40 (±1.24)μg m−3. Dominant biomass burning practices in middle Gangetic regions largely contribute to oxides of nitrogen which on neutralization forms NO3−. Ammonium usually originated from agricultural and vehicular sources (Sharma et al. 2014) found to contribute a significant proportion of WSIC with maximum concentration of 13.21 (±1.25)μg m−3. The concentration of other ions including K+ (5.50±0.51), Mg+2 (1.10±0.31), F− (1.06± 0.42) and Cl− (5.28±0.51) were also found to be at a significant concentration level in airborne PM10. The presence of K+ in the PM10 mass significantly adds to the possibility of local biomass burning as a potent source of aerosols in the region. Ambient fine particulates were found to be associated predominately with sulphate, ammonium and nitrates; relative abundances of which are highly found temporally variable. In


1339

Fig. 10 (a–b) Variation of metal concentration in different size fractions of airborne particulates at ground monitoring station

PM2.5, SO4−2 was reported to have maximum concentration among other WSIC with a concentration of 13.30 ± 1.11 μg m−3 followed by NH4+ (11.54±1.13 μg m−3) and

NO3− (7.71±0.84 μg m−3). Other ions, i.e. K+, Mg+2, F− and Cl− were to found have concentrations of 1.26±0.32, 0.55± 0.11, 0.59±0.13, 3.61±0.19 μg m−3, respectively.

1340


Fig. 11 Variation of WSIC in different size fractions of airborne particulates at ground monitoring station

Chemical composition of particulates Relative chemical composition of particulates in terms of toxic organic fractions, metals and WSIC are presented in Fig. 12a, b. The annually averaged composition of metals (Ca, Fe, Zn, Cu, Pb, Co, Mn, Ni, Cr, Na, K and Cd) has been accounted nearly 14 % of the PM10 mass. Calcium (Ca) constitutes the major fraction of metals contributing 4.3 % of PM10 followed by Na (3.8 %), Fe (1.6 %), Zn (1.3 %) and K (1.2 %). Other trace metals constitute 1.8 % of the PM10 mass. A relatively higher concentration of calcium (Ca) in the PM10 mass depicts the presence of crustal elements originated due to soil erosion or fugitive dust sources. Sharma et al. (2014) reported 17 % major and trace elements in the PM10 mass at New Delhi, situated at the upper Indo-Gangetic plains. Watersoluble ionic components (WSIC) contribute to 26.9 % of the total PM10 with major ions as SO4−2, NO3− and NH4+ with percentage compositions of 8.0, 7.0, and 6.0 %, respectively. Other ions, i.e. (F−, Cl−, K+, Mg+2) constitute 5.8 % of the total mass. Percentage composition of different ions significantly varies with location as well as source. Ram and Sarin (2011) at Kanpur reported an average WSIC contribution of 16 % of the PM10 mass, whereas Sharma et al. (2014) at New Delhi observed 29 % WSIC in it. Variations in ionic composition greatly help in determination of their possible sources. The annually averaged values of ionic constituents comprised of samples having variable concentrations of SO4−2 and NO3−. Within few samples, SO4−2 were reported to be higher,

whereas NO3− has dominated in few ones. Higher SO4−2 values signify the aerosols originated through long-range transport, and higher NO3− in some samples characterizes the dominance of local sources, i.e. biomass burning. Toxic organics associated with particulates was also found to constitute a considerable amount of mass in PM10 (19 %). The contribution of unidentified mass (UM) was estimated by subtracting organic fractions, WSIC and major and trace element concentrations from the particulate mass. About 40 % of the PM10 mass was reported unidentified (UM) which may contain elemental carbon, fly ash, alumino-silicates, carbonate-rich minerals and calcium sulphate. In Fig. 11b, characterization of PM2.5 mass loading represents the various constituents including metals, WSIC and toxic organic fractions. About 11 % of the PM2.5 mass was contributed by metallic species including (Ca, Fe, Zn, Cu, Pb, Co, Mn, Ni, Cr, Na, K and Cd). Calcium represents the maximum proportion of metals in PM2.5 mass (3.2 %) similarly as in PM10. The other important metallic constituents were reported as Fe (0.9 %) and K (0.8 %). The presence of calcium (Ca) and iron (Fe) in higher proportions signifies their crustal origin. Rest metallic content was nearly 6.1 % of the PM2.5 mass comprised of trace metals. The percentage of major ions in PM2.5 follows the trend SO4−2 >NH4+ >NO3− with a total contribution of 25 %. Comparatively, Ram and Sarin (2011) observed nearly 18 % ions in the PM2.5 mass concentration in Kanpur, India. Potassium (K+) is used as a key elemental marker for biomass burning irrespective of


1341

Fig. 12 a Relative composition of PM10 at ground monitoring station. b Relative composition of PM2.5 at ground monitoring station

particle size (Khare and Baruah 2010; Shridhar et al. 2010). Presence of potassium (K+) in airborne particulate (K+:PM2.5, 0.9; PM10, 2.5 %) represents the dominance of biomass burning events in the middle Indo-Gangetic region.

Conclusions Among the very few identified global hotspots of aerosolinduced climate change, Indo-Gangetic plains in South Asia is the most crucial posing threat to nearly 2 billion of loweconomy population. However, information-related particulate mass loading and its compositional variation are

extremely rare in the middle IGP. Therefore, for the present experiment, temporal variation of airborne particulate mass concentration was measured in terms of toxic organics, metals and water-soluble ionic components. The information collected from 1-year (February, 2013 to March, 2014) field experiment was further compared with MODIS aerosol data to identify spatiotemporal trends. Ground-level particulate mass loading (PM2.5: 100.0 ± 29.6 μg/m3 and PM10: 176.1 ± 85.0 μg/m3) was found to exceed the annual permissible limit in a regular basis with 80–100 % of daily concentration persisting well above national standards. Ground-level particulate concentrations express similar behaviour of monsoonal low and winter and summer high. Spatial heterogeneity in AOD was also observed over the IGP both during summer

1342

and winter which may be significantly contributed by transboundary movement of dust storms originated over Thar Desert, Rajasthan and countries from Middle-East which are further aggravated by the development of hygroscopic growth of water-soluble aerosols. Average PM2.5/PM10 ratio indicates significant contribution of finer particulates to major variability of PM10. The concentrations of toxic organics associated with particulates were expressed in terms of BSOF concentrations which also found to contribute a significant proportion (17–19 %) of airborne particulates. Particulate sample was further processed for metals, and trace metals of crustal origin (Ca, Fe, Na and Mg) were found to constitute major fractions of particulates (PM2.5: 4.6; PM10, 9.7 %). Contrastingly, sodium and iron concentrations were mostly prevalent in PM10 (Na: 0.11 to 43.98 μg/m3; Fe: 0.62 to 14.22 μg/m3) rather than finer ones (Na: 0.06 to 9.33 μg/m3; Fe: 0.43 to 7.20 μg/m3). Additionally, trace metals arising from automobile sector, viz. Cu, Pb, Ni and Mn were found below national permissible limits. For both fractions of particulates, SO4−2, NH4+ and NO3− formed the major components (21–23.2 %) of WSIC followed by other ions, i.e. K+, Mg+2, F− and Cl− (5–6 %). Conclusively, it is clearly evident that middle IGP poses a substantial burden of airborne particulates which needs to be further characterized to understand aerosol-climate chemistry of the particular region. Acknowledgments Present research work is financially supported by University Grants Commission, New Delhi (F. No. 41-1111/2012, SR). The MODIS data were acquired from GES-DISC Interactive Online Visualization and Infrastructure (Giovanni) as part of NASA’s Goddard Earth Sciences (GES) Data and Information Services Center (DISC). Authors also acknowledge Director, IESD-BHU; TK Mandal and SK Sharma, National Physical Laboratory-New Delhi for their valuable guidance.

References Aldabe J, Elustondo D, Santamaria C, Lasheras E, Pandolfi M, Alastuey A, Querol X, Santamaría JM (2011) Chemical characterisation and source apportionment of PM2.5 and PM10 at rural, urban and traffic sites in Navarra (North of Spain). Atmos Res 102:191–205 ASTM (2010) ASTM D4600-87 standard test method for the determination of benzene-soluble particulate matter in workplace atmospheres. American Society for Testing and Materials, USA Banerjee T, Singh SB, Srivastava RK (2011a) Development and performance evaluation of statistical models correlating air pollutants and meteorological variables at Pantnagar, India. Atmos Res 99:505– 517 Banerjee T, Barman SC, Srivastava RK (2011b) Application of air pollution dispersion modeling for source-contribution assessment and model performance evaluation at integrated industrial estatePantnagar. Environ Pollut 159:865–875 Begum BA, Biswas SK, Hopke PK (2011) Key issues in controlling air pollutants in Dhaka, Bangladesh. Atmos Environ 45(40:121):11–16 Bullock KR, Duvall RM, Norris GA, McDow SR, Hays MD (2008) Evaluation of the CMB and PMF models using organic molecular

Environ Sci Pollut Res (2015) 22:1329–1343 markers in fine particulate matter collected during the Pittsburgh air quality study. Atmos Environ 42:6897–6904 Ciaccio LL, Rubino RL, Flores J (1974) Composition of organic constituents in breathable airborne particulate matter near a highway. Environ Sci Technol 8:935–942 Continia D, Cesaria D, Genga A, Siciliano M, Ielpo P, Guascito MR, Conte M (2014) Source apportionment of size-segregated atmospheric particles based on the major water-soluble components in Lecce (Italy). Sci Total Environ 472:248–261 Crebelli R, Fuselli S, Conti G, Conti L, Carere A (1991) Mutagenicity spectra in bacterial strains of airborne and engine exhaust particulate extracts. Mutat Res 261:237–248 El-Askary H, Gautam R, Singh RP, Kafatos M (2006) Dust storms detection over the Indo-Gangetic basin using multi sensor data. Adv Space Res 37:728–733 EPA (1999) Inventory of U.S. greenhouse gas emissions and sinks: 1990– 2012. Environment protection agency Fujiwara FG, Gomez DR, Dawidowski L, Perelman P, Faggi A (2011) Metals associated with airborne particulate matter in road dust and tree bark collected in a megacity (Buenos Aires, Argentina). Ecol Indic 11:240–247 Gu J, Pitz M, Schenlle-Kreis J, Diemer J, Reller A, Zimmermann R, Soentgen J, Stoelzel M, Wichmann HE, Peters A, Cyrys J (2011) Source apportionment of ambient particles: comparison of positive matrix factorization analysis applied to particle size distribution and chemical composition data. Atmos Environ 45:1849–1857 Gupta AK, Nag S, Mukhopadhyay UK (2006) Characterisation of PM10, PM2.5 and benzene soluble organic fraction of particulate matter in an urban area of Kolkata, India. Environ Monit Assess 115:205–222 Kar S, Maity JP, Samal AC, Santra SC (2010) Metallic components of traffic induced urban aerosol, their spatial variation, and source apportionment. Environ Monit Assess 168:561–574 Kaufman YJ, Tanre D, Boucher O (2002) A satellite view of aerosols in the climate system. Nature 419:215–223 Keywood MD, Ayers GP, Gras JL, Gillett RW, Cohen DD (1999) Relationships between size segregated mass concentration data and ultrafine particle number concentrations in urban areas. Atmos Environ 33:2907–2913 Khare P, Baruah BP (2010) Elemental characterization and source identification of PM2.5 using multivariate analysis at the suburban site of north-east India. Atmos Res 98:148–162 Liu Y, Franklin M, Kahn R, Koutrakis P (2007) Using aerosol optical thickness to predict ground-level PM2.5 concentrations in the St. Louis area: a comparison between MISR and MODIS. Remote Sens Environ 107:33–44 Mastrangelo G, Fadda E, Marzia V (1996) Polycyclic aromatic hydrocarbons and cancer in man. Environ Health Perspect 104:166–1170 Mouli PC, Mohan SV, Balaram V, Kumar MP, Reddy SJ (2006) A study on trace elemental composition of atmospheric aerosols at a semiarid urban site using ICP-MS technique. Atmos Environ 40:136– 146 Pandey J, Agrawal M, Khanam N, Narayan D, Rao DN (1992) Air pollution concentration in Varanasi, India. Atmos Environ 26B(1): 91–98 Pant P, Harrison RM (2012) Critical review of receptor modelling for particulate matter: case study of India. Atmos Environ 49:1–12 Prajapati SK, Tripathi BD (2008) Assessing the genotoxicity of urban air pollutants in Varanasi City using Tradescantia micronucleus (TradMCN) bioassay. Environ Int 34:1092–1096 Ram K, Sarin MM (2011) Day-night variability of EC, OC, WSOC and inorganic ions in urban environment of Indo-Gangetic plain: implications to secondary aerosol formation. Atmos Environ 45: 460–468 Ramachandran S, Kedia S, Srivastava R (2012) Aerosol optical depth trends over different regions of India. Atmos Environ 49:338–347

Environ Sci Pollut Res (2015) 22:1329–1343 Ramanathan V, Feng Y (2009) Air pollution, greenhouse gases and climate change global and regional perspectives. Atmos Environ 43:37–50 Sharma M, Maloo S (2005) Assessment of ambient air PM10 and PM2.5 and characterization of PM10 in the city of Kanpur, India. Atmos Environ 39:6015–6026 Sharma SK, Kumar M, Rohtash GNC, Saraswati SM, Mandal TK (2014) Characteristics of ambient ammonia over Delhi, India. Meteorog Atmos Phys 124:67–82 Shridhar V, Khillare PS, Agarwal T, Ray S (2010) Metallic species in ambient particulate matter at rural and urban location of Delhi. J Hazard Mater 175:600–607

1343 Viana M, Pandolfia M, Minguillo MC, Querola X, Alastueya A, Monfortb E, Celadesb I (2008) Inter-comparison of receptor models for PM source apportionment case study in an industrial area. Atmos Environ 42:3820–3832 Wang J, Christopher SA (2003) Intercomparison between satellite-derived aerosol optical thickness and PM2.5 mass implications for air quality studies. Geophys Res Lett 30: 1944–8007 WHO (2013). Global burden of disease 2013. World health organization Willeke K, Baron PA (1993) Aerosol measurement. Van Nostrand Publisher, New York