Atmospheric Bulk Deposition of PAHs over Brahmaputra Valley : Characteristics and Influence of Meteorology

Bulk atmospheric deposition of Polycyclic Aromatic Hydrocarbons (PAHs) in Guwahati city of the Brahmaputra Valley have been characterised for a period of one year. The ∑PAHs (USEPA’s priority 16) and benzo(s)pyrene (BaP) concentrations in the collected bulk deposit ranged between 2.2 and 1035 ng mL, and BDL and 5.6 ng mL respectively. Greater deposition of PAHs was observed during the dry season and the deposition of low molecular weight PAHs (LMWPAHs) were particularly high. The study revealed explicit effect of the prevailing meteorology on the bulk atmospheric deposition characteristics of PAHs. Greater number of inversion days and lowering of inversion height during the dry season were found to enhance deposition of PAHs. Diagnostic ratios indicated pyrogenic origin of the PAHs derived from combustion of diesel, coal and wood combustion. The deposition of LMWPAHs was greatly influenced by temperature. The diurnal variation of relative humidity, and wind pattern were found to affect deposition of PAHs.


INTRODUCTION
Polycyclic Aromatic Hydrocarbons (PAHs) make a group of ubiquitous pollutants of the atmosphere, especially in the urban/industrial regions due to their intense source strengths (Brown et al., 1996;Nielsen et al., 1996;Coleman et al., 1997;Simcik et al., 1997;Lim et al., 1999;Birgül et al., 2011).The PAHs have gained much attention due to their mutagenic and carcinogenic potentials.As many as seven members of the list of the 16 priority PAHs of the USEPA are probable human carcinogens (Baek et al., 1991;Petry et al., 1996) and Benzo(a)Pyrene (BaP) serves as an indicator of the whole group of PAHs for its reputation of being a carcinogen.Higher health risk associated with the non-dietary PAHs exposure via dust (Wang et al., 2013) and due to their multiple sources unlike other persistent organic pollutants, they cannot be curbed by introducing substitute chemicals (Jang et al., 2013).
Distribution and concentration of PAHs in the environment including air, river water, lakes, sediments and vegetation have been extensively studied (e.g., Patrolecco et al., 2010;Navarro-Ortega et al., 2010;Kafizadeh et al., 2011;Liu et al., 2012;Yan et al., 2012;Slezakova et al., 2013;Hussain et al., 2014;Wu et al., 2014;Hussain and Hoque, 2015;Hussain et al., 2015Hussain et al., , 2015a;;Cheruiyot et al., 2015).Atmospheric deposition of PAHs has not attained much focus compared to other environmental matrices.Atmospheric deposition of contaminants is a determining factor to total load of pollutant in various aquatic and terrestrial ecosystems (Mai et al., 2003).Yet studies related to atmospheric bulk deposition of PAHs are scanty and limited data have been reported.Majority of studies on atmospheric deposition found in the literature are aspired to perform wet only or dry deposition flux of PAHs.In order to estimate atmospheric deposition loadings it is important to examine the bulk composition.Though some studies (e.g., Manoli et al., 2000;Grynkiewicz et al., 2002;Garban et al., 2002;Montelay-Massei et al., 2003;Delhomme et al., 2007;Montelay-Massei et al., 2007;Esen et al., 2008;Wang et al., 2011) reported various attributes of bulk deposition of PAHs, there has been no such study conducted yet for the South Asian region.
Therefore, this study was taken up to understand various attributes of atmospheric bulk deposition of USEPA's 16 PAHs with a focus to appreciate the spatial and seasonal deposition flux variabilities, composition profile and sources.We made an attempt to understand the influence of the prevailing meteorology on the bulk attributes of the PAHs in the atmospheric process of bulk deposition.

Study Area
The study was carried out in Guwahati city of the Brahmaputra Valley of the Northeast India.The city is situated between the Brahmaputra River and the foothills of Shillong plateau, geographically positioned around 26° 10'0"N and 92°49'0"E and spread over an approximate area of 340 km 2 .Guwahati has bowl shaped topography and an undulating terrain, with its altitude varying from 49.5 to 55.5m above mean sea level.
Guwahati has a warm and humid climate, and the monsoon wind brings the major share of annual rainfall during the summer month.The sampling schedule with their respective mean monthly total rainfall (mm), mean monthly temperature (°C) and mean monthly relative humidity (%) of the study period has been put up in Table S1.The region represented in this paper is developing fast and biomass burning is common in rural areas and during festivities (Deka andHoque, 2014, 2015).Biomass burning releases large volume of PAHs.In a study Kamal et al. (2016) reported that about 88% of PAHs originated from wood combustion in Punjab Province of Pakistan.
As per the 2011 Government of India census, Guwahati recorded a population of ~1.0 million at a decadal growth rate of 8%.Over the last decade Guwahati has witnessed a vehicular growth (both light and heavy vehicle) of around 87%.At the end of the first quarter of 2013 there were 58,638 vehicles registered in Guwahati.
Five representative locations were chosen for collecting deposition samples on the basis of landuse of Guwahati city viz.industrial (Noonmati-Narengi area; S1), commercial (Machkhowa-Fancy Bazar; S2), high traffic roadside (Gauhati University; S3), residential (Khanapara-Beltola; S4) and forest site (Basistha; S5).Sampling sites of Guwahati city has been illustrated in Fig. 1.The site S1 hosts an oil refinery and several units of carbon producing industries, and S2 is a commercial hub together with a busy bus station.The site S3, though located in a University campus, is near a busy traffic junction.

Sampling
Atmospheric bulk deposition sampling was carried out during February 2011 to February 2012.The study period was classified into dry and wet seasons; seven months of the sampling year had more than 10 rainy days, which were considered as wet periods.
Thus, the sampling period from March/April-2011 to September/October-2011 were classified as wet season.Monthly deposition samples were collected by a passive sampler which is made of a stainless steel funnel fitted into a collector bottle (amber glass bottles).The collector bottles were pre-washed with acetone and dried before installing in an isothermal box made up of polystyrene to withstand temperature variation, and avert evaporation.
Triplicate samplers were kept open to grab the atmospheric fall-out for a continuous period of about 4 weeks at an average height of 7 m.In addition, a 'field blank' sampler covered with aluminium sheet was kept for blank correction.
Three monthly samples of each site were mixed to make a composite sample, which was stored at 4°C till further analysis.
Overflow occurred at 5 instances during the sampling periods of May to September 2011.For these samples, maximum capacity of collecting bottles was taken into consideration while computing the results.

Sample Extraction and Clean up
The PAHs were recovered from samples by liquid-liquid extraction (organic-aqueous partitioning) in a separating funnel.One litre of each sample was extracted three times with a mixture of hexane and dichloromethane (v/v, 85/15) (Ollivon et al., 1999).All 3 extracts were combined and reduced to about 2 mL, which was then exchanged to cyclohexane for clean up.
Clean up was done, as per USEPA method (Method no.3630C) by a silica column packed with 10 gm of activated silica of 100-200 mesh (Merck).The eluate containing desired PAHs was reduced in a rotary evaporator and adjusted to 1 mL in acetonitrile.

Analysis
Analyses were done in High Performance Liquid Chromatography (HPLC) (Waters) equipped with Ultraviolet (UV) detector (W22489) and C18 column (4.6 mm × 250 mm, 5 µm-particle size) in a gradient mode -acetonitrile/ water.Quantifications of the PAHs were done by calibrating with both external and internal calibration standards.Compound concentrations were calculated using both peak area response and mean relative response factors (RRF) (Hussain et al., 2015)

Quality Control
All samples and blanks were spiked with the mixture of four internal standards prior extraction to determine analytical recovery efficiencies.Average recoveries of the internal standards were found to be 68 ± 27% for naphthalene-d8, 67 ± 1% for phenanthrene-d10 and 101 ± 16% for chrysene-d12.For statistical analysis of the data set below detection limit (BDL) values were replaced by half of LOD values.The LODs were calculated as two times the standard deviation of the replicate analysis of the lowest standard (n = 5) (EPA17-A).Analysis of field blank samples was performed to avoid contamination.
Sampling of semivolatile organic compounds with similar samplers were also reported in literature (Dickhut and Gustafson, 1995;Halsall et al., 1997;Manoli et al., 2000; Fig. 1.Map of the study area -Guwahati and the surrounding region of Asia.Garban et al., 2002;Ollivon et al., 2002;Esen et al., 2008).Possible artefacts related to this type of sampler would be (i) degradation of PAHs at metal surface due to atmospheric radicals, solar radiation etc., (ii) re-volatilization of PAHs due to up-heating of the metal, (iii) particle blow-off due to high wind speed (Esen et al., 2008).

Meteorological Conditions
The summary of weather data procured from the nearest Meteorological Centre at the LGB International Airport, Guwahati is shown in Table S1.There were 231 rainy days during the sampling period with total precipitation of 1367 mm.Guwahati had average rainfall of 1700 mm in over last ten years.The mean monthly temperature ranged between 16.6°C and 29.5°C during the study period, and mean monthly relative humidity (RH) ranged between 15.3 and 27.6%.PAHs are often examined with varying meteorological conditions and significant statistical relationships are observed between the two in many cases (Barrado et al., 2013;Masiol et al., 2013).

PAHs in Bulk Deposition
The concentrations of PAHs in the bulk deposit during 2011-2012 have been put up in Table 1.The concentrations of ∑PAHs in the bulk deposit of Guwahati were found to be between 2.2 and 1035 ng mL -1 .The BaP concentrations varied from BDL to 5.6 ng mL -1 with mean of 0.2 ± 1 ng mL -1 .The percentage contribution of BaP in ∑PAHs was found to be 0.3%.Individually, Naph followed by BaA, Chry, Flan and Phen were found to be more abundant in the bulk collected deposits over Guwahati.Manoli et al. (2000), in a previous study, reported the dominant contribution of Naph in atmospheric bulk deposition samples.They attributed that high water solubility of Naph could be the reason for such high abundance.But water solubility may not be the only reason for the dominance of Naph in bulk samples as despite of their low water solubility, BaA and Chry were also found to be the dominant Table 1.Mean ± SD (min-max) PAHs in bulk deposition*.Despite there being differences in the type of deposition (exclusively wet or bulk), sampling, extraction and analytical methods and number of PAHs considered in various studies in literature, levels of ∑PAHs and BaP of present study were compared with previous works (Table 2) which indicates higher levels of PAHs in the present study.

Bulk Deposition Mass Flux
The flux of ∑PAHs in Guwahati city has been illustrated in Fig. 2. The flux of ∑PAHs ranged between 63.5 and 3270 µg m -2 .The mean ∑PAHs flux was found to be maximum at S2 and minimum at site S3.The deposition rate of BaP ranged from BDL to 11.4 µg m -2 .The daily ∑PAHs flux varied from 2.1-401.4µg m -2 day -1 and the flux of BaP varied from BDL-0.24 µg m -2 day -1 .The flux of PAHs of present study were on much higher side than that of highly industrial cities like Manchester and Cardiff (UK) (1-24.2 and 8-19.6 µg m -2 day -1 respectively as reported by Halsall et al. (1997); the fluxes of Paris (0.17-2.µg m -2 day -1 ) reported by Ollivon et al. (2002); Northern Greece (0.21-1.1 µg m -2 day -1 ) as reported by Manoli et al. (2000); Southern Germany (0.2-1.06 µg m -2 day -1 ) as reported by Gocht et al. (2007); Hungary (1.04-13.73µg m -2 day -1 ) as reported by Kiss et al. (2001); in New Jersey, Mid-Atlantic east coast region (0.0004-3 to 0.14 µg m -2 day -1 ) as reported by Gigliotti et al. (2005); Sweden (µg m -2 day -1 ) as reported by Brorström-Ludén et al. (1994), Beijing, North China (1.7 ± 1.09 to 14.25 ± 12.33 µg m -2 day -1 ) as reported by Wang et al. (2011) and Shanghai, China (1.6 µg m -2 day -1 ) as reported by Yan et al. (2012).Bowl shaped topography of the study area renders inversion conditions from the evening hours throught the night, which essentially creates condition for high deposition of pollutants.This could be the principal reason which may be attributed to this high daily flux of PAHs in the study area.
The annual flux of PAHs in the entire city of Guwahati was calculated as per the method adopted by Yan et al. (2012): where, R ̅ is the annual rainfall of Guwahati in the last ten years (1007.2mm) and S t is the area of Guwahati city (340 km 2 ).
where C i is the concentration (ng mL -1 ) of PAH of each event, R i is the total rainfall of each event (mm).where, V t is sampling amount of event (ml) and S b is the area of the funnel.The calculated mean annual flux of PAHs of Guwahati city was found to be 4558kg.This was found to be greater than that of Shanghai, China (4148kg yr -1 ) as estimated by Yan et al. (2012).

Spatial and Seasonal Variation of PAHs
The spatial variations were also investigated on the basis of landuse pattern.The maximum ∑PAHs concentration was observed at the industrial site with mean value of 157 ± 277 ng mL -1 followed by residential site of 76 ± 101 ng mL -1 , commercial site of 54 ± 51 ng mL -1 , forest site of 34.8 ± 25 ng mL -1 and traffic site of 34.5 ± 29 ng mL -1 .The pattern of BaP concentrations follows the order: industrial site > residential site > forest site > traffic site > commercial site.Yet the BaP concentrations were observed to be almost in the same range in all sites except at the industrial site with very high concentration level (0.7 ± 1.8 ng mL -1 ).
The maximum ∑PAHs concentration at S1 could be attributed to major industrial activities.However, residential site with such high concentration could be attributed to highway traffic which is not so far from the sampling area.Similar PAH concentration levels of forest site with that of traffic site is another major concern.High amounts of wet and dry deposition of PAHs in remnants of urban forest was also reported by Wong et al. (2004).
The ∑PAHs concentration in bulk deposition was also found to vary significantly between dry and wet seasons during the study period (Fig. 3(a)).During wet season, the bulk deposition of ∑PAHs concentrations ranged between 14.1 and to 100.2 ng mL -1 with mean of 43.9 ± 32 ng mL -1 (Table 3).However, ∑PAHs concentrations varied from 25.5 to 316.4 ng mL -1 with mean of 93 ± 101 ng mL -1 during dry season.Individually most of the PAHs were also observed with increased concentrations during dry season compared with that of wet season except for BaA.
The observed pattern of concentrations could be attributed to the prevailing atmospheric conditions.From the analysis of air mass back trajectories computed from NOAA's HYSPLIT model (Fig. 3 From the number of inversion days and mean inversion height during the entire study period, it may be noted that there was an increase in the number of inversion days during dry season in comparison with the wet season (Fig. 3(c)).It is pertinent to note that there was a lowering of mean inversion height by approx.100-200 m with the increase in the number of inversion days during dry season that may lead to more accumulation of pollutants during the season.Also, in a previous study maximum of PCB dry deposition fluxes occurred in winter (Tseng et al., 2014).
The wash out of pollutants by heavy rainfall could also be the possible cause of lower concentration during wet season of the year.Moreover, the degradation rate of PAHs is usually faster under high temperature and strong solar radiation during the period.Further, during wet season in  India wind from sea -the monsoon wind during the summer months -also dilute local pollutants in air (Gu et al., 2010).Similar temporal trend for bulk deposition of PAHs have also been reported by other researchers (e.g., Smith and Harrison, 1996;Manoli et al., 2000;Ollivon et al., 2002).Variation of ∑PAHs concentrations on the basis of their molecular weight (HMWPAHs and LMWPAHs) during the study period was also examined (Fig. S1).For both LMWPAHs and HMWPAHs maximum concentration peak was observed during November-December with very high contribution of LMWPAHs to ∑PAHs.In addition to maximum ∑PAHs concentration during Nov/Dec 2011 two minor concentration peaks were also detected during the sampling period of May/June 2011 and July 2011.
In a study, Zhang et al. (2016) stated that seasonality of ∑PAHs deposition, which could be an indicator of particulate deposition of an area.

PAH Profiles
The profiles of PAHs were dominated by 2 and 4-ring members of PAHs (Fig. 4).Dominance of PAHs profile by LMWPAHs could be attributed to their higher level of solubility.The 'multi-hop' nature of LMWPAHs also made them more susceptible to long range transport and continuous cycle of deposition and re-evaporation (Agarwal, 2009).The low molecular weight PAHs found to have significant acute toxicity (Yang et al., 1991), which is a major concern.Lower representation of 5 and 6-ring PAHs might be due to their low solubility in wet fraction of deposits.Congener profiles as per USEPA recommended 7 carcinogenic PAHs (CPAHs) viz.BaA, BaP, BbF, BkF, Chry, DBA and IP, and non-carcinogenic PAHs are shown in Fig. 4. The CPAHs contribution in bulk deposition was found to be 31%.Among the carcinogenic lot the representation of BaA and Chry were found to be maximum in all sites.

Meteorology and Deposition of PAHs
Transmissions and deposition of pollutants are greatly affected by prevailing meteorology (Suryani et al., 2015).Table 4 illustrates significant correlation of PAHs with   various meteorological parameters.It was observed that Naph and Phen showed significant negative correlation with mean temperature, while BbF and DBA displayed significant negative correlation with mean humidity.Naph, Acy, Phen and ∑PAHs exhibited good negative correlation with mean wind speed, while Acy, Acen, BaA, BaP and ∑PAHs showed significant negative correlation with wind direction.Thus LMWPAHs showed negative correlation with mean temperature while HMWPAHs with relative humidity.
Individual PAH compounds behave differentially under different meteorological and climatic conditions depending upon their molecular weight and structure.It has been reported that temperature has significant impact on the dispersion of certain category of PAH compounds having low molecular weight (Pankow et al., 1993).According to Stronguilo et al. (1994) LMWPAHs with higher vapour  pressure and higher level of solubility become more susceptible to volatilization at room temperature (20°C).Yamasaki et al. (1982) and Pankow et al. (1993) have also reported that meteorological parameters, especially, temperature and relative humidity play a significant role in dissipation of PAHs.Pankow et al. (1993) also observed that gas/particle partitioning process is negatively dependent on relative humidity over the range of 40% ≤ RH ≤ 95%, for some PAHs and groups of PAHs in case of urban particulate matter.
Significant negative relationship of PAHs with wind speed was an indicator of high dispersion of atmospheric PAHs with increasing wind speed.The negative correlation of PAHs with wind direction was an evidence of deposition of PAHs mainly through northerly or north-easterly wind, especially during dry season of the year in the study area.As the wind direction was shifting towards southerly or south-westerly which was mainly prevalent during wet season of the sampling year, deposition of PAHs was found to be decreased.This phenomenon has been illustrated in Fig. S2.
Based on these observations, relationship of PAHs with temperature and relative humidity was established.Relationship of logarithmic ∑PAHs, ∑LMWPAHs and ∑HMWPAHs with temperature [maximum temperature (T max ), mean temperature (T mean ), diurnal temperature range (DTR)], relative humidity [maximum relative humidity (RH max ), mean relative humidity (RH mean ) and diurnal relative humidity range (DHR)] have been plotted, which is shown in Figs.5(a), 5(b) and 5(c) respectively.∑PAHs and ∑LMWPAHs showed significant negative relationship with T max (r 2 = 0.32 and r 2 = 0.36 respectively) and T mean (r 2 = 0.4 and r 2 = 0.4 respectively).However, positive relationship was observed for DTR with r 2 of 0.36 and 0.24 for ∑PAHs and ∑LMWPAHs respectively.Positive relationship was also found for both ∑PAHs and ∑LMWPAHs with DHR (r 2 = 0.3 and 0.32 respectively).However, in the case of HMWPAHs no such relationship has been observed with T max , T mean , RH max and RH mean .Instead, HMWPAHs was found to exhibit positive relationship DTR and DHR (r 2 = 0.22 and 0.20 respectively).
From the above relationship it is established that the concentration of LMWPAHs in the atmospheric deposition was greatly influenced by temperature, which is not the case for HMWPAHs.At ambient air temperature, the LMWPAHs (2-to 4-ring PAHs) mostly occur in vapour phase while the HMWPAHs (5 or more rings) are associated with particulate phase (Arey and Atkinson, 2003).So, the LMWPAHs with high vapour pressure are more prone to volatilization than HMWPAHs (Stronguilo et al., 1994;Huesemann et al., 1995;Mackay and Calcott, 1998).Vapour pressure controlling volatilization has been found to be regulated by temperature, humidity and wind speed.The volatilization of PAHs increases from soil, atmospheric particles, water and vegetation with rise of temperature (Ausma et al., 2001;Sofuoglu et al., 2001).Thus, volatilization of LMWPAHs is more dependent upon air temperature than HMWPAHs.According to Sofuoglu et al. (2001) temperature deviation accounted for gas phase variability (23-49%) in the case of Acen and Chry, yet only 1-6% for HMWPAHs.Thus, the relationship of LMWPAHs with temperature in the present study could be attributed to temperature dependent scavenging process for the LMWPAHs as documented by Ollivon et al. (2002).Moreover, the maximum temperature values were found to have greater influence on LMWPAHs.Hassan and Khoder (2012) in their study, also, reported similar trend of decrease in total PAHs concentration in particulate and gaseous phase with increasing temperature.Tsapakis and Stephanow (2005), Cincinelli et al. (2007) also showed similar studies.
Present study also revealed an increase of PAHs concentration with increase of diurnal variation of temperature.The LMWPAHs are often found to be influenced by seasonal and diurnal changes with long range transport and deposition on a surface through continuous air-surface exchange by means of succeeded cycles of deposition and   re-evaporation as reported by Yamasaki et al. (1982), Hoff and Chan (1987), Keller and Bidleman (1984).Greater gaseous state of LMWPAHs was also reported due to increase in temperature in summer, contrary to winter, which increases vapour pressure of LMWPAHs (Kaupp and McLachlan, 1998;Odabasi et al., 1999).This could be applicable in the case of high diurnal temperature deviations also.Murray et al. (1974) showed that for a temperature change of 20°C there is approximately an order of magnitude change in the vapour pressure of PAHs.
Positive effect of diurnal variation of relative humidity was also found on PAHs concentrations.Relative humidity in the study area was found to rise up to 100%.Thibodeaux et al. (1991) reported an increase in absorption of semi volatile organic compounds as relative humidity approaches 100%.Such increase at very high relative humidity is due to partitioning into the liquid or nearly-liquid phase when the gas phase is nearly saturated with water.

CONCLUSIONS
The study revealed that considerable volume of PAHs gets deposited over the city under study.The 2-and 4-ring PAHs were mostly dominated in the deposited PAHs.Spatial and temporal variations of PAH concentrations were observed in bulk deposition samples.Representation of USEPA's 7 carcinogenic PAHs was found to be significant mainly in industrial site.Diagnostic ratios of the PAHs supported for pyrogenic inputs of the atmospheric PAHs.
The prevailing meteorology has significant role in the bulk deposition of PAHs -both in terms of the flux and the ring size of the PAHs.The deposition of LMWPAHs was greatly influenced by temperature, which is not the case for HMWPAHs.Positive effect of diurnal variation of relative humidity was also found on bulk deposition of PAHs.And, significant negative relationship of PAHs with wind speed was an indicator of high dispersion of atmospheric PAHs with increasing wind speed.
(b) (i)) and the vertical temperature profile of the study area (Fig. 3(b) (ii)) computed form the data obtained from Wyoming university atmospheric data repositories, it is understood that the prevailing atmospheric condition with low wind, subsidence and inversion conditions were the guiding factors to high PAHs accumulation during the period of Nov/Dec 2011.During the period of May/June 2011 and Jul 2011, atmospheric dispersion was very low (Fig. 3(b) (iii) and (v)) due to the development of temperature inversion (Fig. 3(b) (iv) and (vi)) in the lower atmosphere, which are the factors of accumulation of PAHs.

Fig. 3
Fig. 3(a).Comparison of PAH concentrations during dry to wet periods.

Fig. 3
Fig. 3(c).Mean inversion height (m)Vs number of inversion days during study period.

Fig. 4 .
Fig. 4. The profile of the PAHs in the bulk deposit: 2-and 4-ring compounds are more abundant, and CPAHs make 31% of the bulk deposition of PAHs in Guwahati city.
Zhang et al. (2011)an and Phen in bulk deposition samples.Zhang et al. (2011)reported equal distribution of Naph, BaA and Chry both in gaseous and particulate phase PAHs.Association of these compounds in both the phases in atmosphere could possibly attribute higher concentration of Naph, BaA and Chry in bulk deposition. compounds

Table 2 .
PAHs concentration (ng mL -1 ) in bulk precipitation around the world.

Table 3 .
Mean and range of PAHs (ng mL -1 ) in bulk deposition during dry and wet season.

Table 4 .
Correlations of PAHs with various meteorological parameters.