Transportation Research Part D 48 (2016) 113

Contents lists available at ScienceDirect

Transportation Research Part D

journal homepage: www.elsevier.com/locate/trd

Characteristics of air pollutants at near and far field regions

of a national highway located at an industrial complex
S.M. Shiva Nagendra , M. Diya, V.S. Chithra, Jyothi S. Menon, Anju Elizbath Peter
Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600036, India

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents the characterization of air quality monitored at near field region (NFR)
Available online 17 July 2016 and far field region (FFR) of a national highway located at an industrial complex. The pol-
lutants such as PM10, SO2 and NO2 were monitored in two campaigns (11th September to
Keywords: 18th October 2012 and 18th January to 17th February 2013). The 24 h average PM10 con-
Dispersion centration at NFR and FFR were found to be 86.69 18.56 lg/m3; 73.16 16.21 lg/m3 and
Far field region 89.44 18.69 lg/m3; 81.91 16.42 lg/m3, respectively during first and second campaign.
Highway
In both the campaigns PM10, SO2 and NO2 concentration at NFR was higher than FFR.
Industrial complex
Near field region
The chemical characterization of PM10 at NFR and FFR indicated the abundance of major
Stack emission elements such as Na (NFR = 30% and FFR = 32%), Ca (NFR = 12% and FFR = 14%) and ions
Secondary pollutants namely NO
+
3 (NFR = 71% and FFR = 68%) and NH3 (NFR = 15% and FFR = 19%). Further, at
2


Heterogeneous traffic FFR, SO4 and NO3 were found to be 18% and 35% higher than NFR indicating the conver-
sions of SO2 and NO2 concentration into secondary particles. The measured SO2 and NO2
concentrations were 23 and 21% lower at FFR when compared to NFR confirms the sec-
ondary formation.
The CALPUFF, EPA regulatory model was set up to understand the dynamics of air pollu-
tants at the industrial complex. The predicted PM10, SO2 and NO2 concentrations at NFR
and FFR were found to be 32.31 1.56 lg/m3 and 31.35 1.27 lg/m3; 0.37 0.21 lg/m3
and 0.06 0.04 lg/m3; 12.83 6.55 lg/m3 and 4.67 2.77 lg/m3, respectively. The model
showed moderate predictions for PM10 (R2 = 0.440.52), SO2 (R2 = 0.410.51) and NO2
(R2 = 0.450.61) concentrations.
2016 Elsevier Ltd. All rights reserved.

1. Introduction

Air pollution remains a major environmental problem in the industrial complex despite of various efforts in many devel-
oping countries. In the recent past, emissions from the industrial sources are showing decreasing trend, the emissions from
vehicles at industrial complex has increased due to increase in vehicle operations and economic growth. Large quantities of
chemical compounds are emitted into atmosphere as result of anthropogenic processes. These emissions lead to a complex
array of chemical and physical transformations resulting in apparently diverse effects as photochemical air pollution, acid
deposition, long-range transport of chemicals and global weather modifications (Atkinson, 1990). Several stringent actions
were in developed countries to tackle the air pollution problem in the industrial area. In developing countries, air quality
management at the industrial area is still a challenge. A study carried out by the central pollution control board (CPCB) in

Corresponding author.
E-mail address: snagendra@iitm.ac.in (S.M. Shiva Nagendra).

http://dx.doi.org/10.1016/j.trd.2016.07.006
1361-9209/ 2016 Elsevier Ltd. All rights reserved.
2 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113

India showed that 43 industrial clusters are having comprehensive environmental pollution index (CEPI) greater than 70 (on
a scale of 0100) and have been identified as critically polluted (CPCB, 2009a,b). In industrial complex, motor vehicles also
contributes considerable amount of pollution, because most of the industrial areas are connected to national highways for
the purpose of quick transport of raw materials and finished products. In general, traffic on these roads is dominated by
heavy duty diesel vehicles. Highways having increased heavy duty vehicles traffic showed elevated particle matter concen-
trations (Ntziachristos et al., 2007). At present, chemical characterization of PM at industrial area is scanty. The chemical
composition at near field and far filed regions and their associations with different emissions sources at near filed and far
are not well understood. The information on PM chemical characterization is important to study the effects on human health
and surrounding environment.
Dispersion models are widely used to understand the dynamics of air pollutants in near and far field regions. But most of
the dispersion models (AERMOD, ADMS, ISCST3, CALINE4, etc.) often assume steady state, straight-line transport of pollu-
tants in time and space. In a complex terrain, this assumption is inappropriate where the wind field is inhomogeneous.
The U.S. Environmental Protection Agency (USEPA) has recommended CALPUFF Lagrangian puff model (Earth Tech, Concord,
MA) for long range transport modeling (USEPA, 2000), because of its ability to handle complex three-dimensional wind
fields. CALPUFF model characteristics include, the capability to treat time-varying point, area, volume and buoyant line
sources; suitability for modeling domains from tens of meters to hundreds of kilometers from a source or source complex;
predictions for averaging times ranging from one-hour to one year; applicability for inert pollutants and those subject to lin-
ear removal and chemical conversion mechanisms; and applicability for rough or complex terrain situations and overwater
scenarios (USEPA, 1998). According to USEPA, the CALPUFF may be considered for near field (less than 50 km) regulatory
applications when assumptions of steady state straight-line transport in time and space are inappropriate (USEPA, 2008).
Near field applications of CALPUFF has received little attention, especially in settings with complex terrain and wind.
In the present study, an attempt was made to characterize the air pollutants at near field region (NFR) and far field region
(FFR) of a national highway passing through an industrial complex. The dispersion of criteria pollutants emitted from the
industrial complex was also investigated.

2. Methodology

2.1. Study area

The study area-Sriperumbudur (latitude 12.9700N and longitude 79.9500E) is in the state of Tamil Nadu, India, and
located 40 km west south-west of Chennai city. Sriperumbudur is strategically placed on the Chennai-Bangalore national
highway four (NH4). Physiographically, the study area terrain is complex and can be classified into plain land with low-
lying hills. The general elevation of the study area ranges from 20 to 140 m above mean sea level (MSL). The elevation con-
tour indicates the area is sloping towards east. Climate of the study area is semi-arid. Wind direction is mostly southwest,
west and northwest with an average speed of 7 kmph. The temperature throughout the year is ranging from 37.5 to 20.5 C
and relative humidity is from 60 to 74%. The study area receives good amount of rainfall during northeast monsoon.
There are more than 50 small and large industrial firms are functioning in the state industries promotion corporation of
Tamil Nadu limited (SIPCOT), industrial park at Irrungattukottai, Sriperumbudur. Major industrial companies working in this
area are Hyundai Motors, Hindustan Motors, Mitsubishi, BMW, Foxconn, Flextronics, Jabil Green Point, Sanmina, Aspocomp,
Nokia and Nokia Siemens, Samsung, Nissan, Ford, Renault, Laird Group Plc, Dell, Mahindra & Mahindra, etc. Hyundai was one
of the initial investors when they established their Indian car operations in 1999 in Sriperumbudur. The major contributors
of pollution in this area are industries and vehicles. Unlike urban roads, which are having large fraction of light duty vehicles,
the Chennai Bangalore national highway has comprised of equal distribution of heavy duty and light duty vehicles. PM10, SO2
and NO2 measuring instruments have been setup at distance of 10 m (NFR) and 200 m (FFR) from the centerline of the NH4
(Fig. 1). There are 5 stacks present within a radius of 10 km to the monitoring station.

2.2. Air quality monitoring

Two monitoring campaigns were conducted for the measurement of PM10, SO2 and NO2 during 11th September to 18th
October 2012 (S1) and 18th January to 17th February 2013 (S2) at both NFR and FFR of Chennai Bangalore national highway.
High volume samplers (Envirotech APM 460NL) with gaseous attachment (APM 411TE) were used to measure PM10, SO2 and
NO2 concentrations. The 24 h average PM10 mass was collected on glass microfiber filters (Nupore Filtration Systems Private
Limited, India) of size 20.3  25.4 cm with an airflow rate of 1.1 m3/min. About 38 samples were collected per campaign by
covering all days of the week to give a fair representation. The instruments were kept at a height of 1.5 m from the ground
level. Glass fiber filters were weighted twice before and after sampling by a microbalance with a sensitivity of 1 lg. Before
weighing filters were equilibrated in a desiccator at room temperature with a relative humidity of 4050% for 24 h. To reduce
gravimetric bias due to filter handling during and after sampling, field and laboratory blank filters were collected and ana-
lyzed. Filters were placed in aluminum foil sealed cassette while carrying from the field. From the weight differences and
airflow rate, the PM10 concentrations (lg/m3) were determined. Further, the weighed filter papers were preserved in a free-
zer until further chemical analysis.
 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113 3

Fig. 1. Monitoring site and location of sampling stations at Chennai-Bangalore national highway. Monitoring station 1 (NFR) and monitoring station 2 (FFR).

The 24-h average SO2 and NO2 levels in ambient air were monitored using the gaseous attachment in the sampler with an
airflow rate of 0.5 L/min. Modified West and Gaeke method was used for SO2 analysis and Jacobs and Hochheiser method for
NO2 analysis (CPCB, 2011). The SO2 from air was absorbed in a solution of potassium tetrachloromercurate (TCM). A
dichlorosulphitomercurate complex, formed was made to react with para-rosaniline and formaldehyde to form the intensely
coloured pararosaniline methylsulfonic acid. The absorbance of the solution was measured by means of a UVVisible spec-
trophotometer at 560 nm. NO2 was collected by bubbling air through a solution of sodium hydroxide and sodium arsenite.
The concentration of nitrite ion (NO
2 ) produced during sampling was determined colorimetrically by reacting the nitrite ion
with phosphoric acid, sulfanilamide, and N-(1-naphthyl)-ethylenediamine di-hydrochloride (NEDA) and measured the
absorbance of the highly coloured azo-dye at 540 nm.
In the present work, PM10, PM2.5 and PM1 concentrations were also monitored for 24 days (11/9/2012 to 6/10/2012) at the
study region using the Environmental Dust Monitor (GRIMM-107, Make: GRIMM Aerosol Technik). GRIMM dust monitors
samples air at the rate of 1.2 L/min with a lower cut-off at 0.2 lm. The instrument was kept at a distance of about 10 m from
the national highway and the sampling inlet was placed at 1.5 m above the ground level. The PTFE Filter of diameter 47 mm
and pore size 0.2 lm was used for PM sampling. Filters were changed once in 7-days during the monitoring period. In order
to achieve the data quality control instrument was calibrated using standard calibration procedure. The meteorological
parameters such as wind speed, wind direction, temperature, humidity, pressure and precipitation at the study region were
collected from meteorological station set up at Hyundai motors for the study period (11 September 2012 to 20 February
2013) for which the pollutants concentration were measured.

2.3. PM chemical analysis

Chemical characterization of PM was conducted only during first monitoring campaign. For PM chemical characterization,
glass fiber filter papers were divided into two halves. One half was used for ion analysis and the other half was used for ele-
mental analysis. Extraction and analysis of elements were carried out as per the USEPA Compendium Methods (USEPA,
1999a, 1999b). A hot-plate digestion method was used for the extraction of elements from the first half of filter. Digested
samples are filtered to remove any solid residues and refrigerated for further analysis. The extracted sample was then
injected into an inductively coupled plasma with optical emission spectroscopy (ICP-OES, make: Optima 5300-DV ICP-
OES) for analyzing the elements viz. Al, As, B, Ba, Ca, Cd, Co, Cr, Cu, Fe, Ga, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Rb, Se, Sr, Ti, V
and Zn. Out of 25 species, 15 elements were found to be present in the PM samples and other elements were in the range
of below detection limit of the instrument.
The other half of the filter paper was extracted separately for ion analysis by using standard operating procedures (SOP)
given by CPCB (2010). Extracts were then injected into an ion chromatograph (Dionex ICS-2500) for analyzing the water sol-
uble viz. Na+, NH4+, K+, Ca2+, Mg2+, F
, Cl
, NO

2

2 , NO3 and SO4 . A Dionex AS11 separation column with an ASRS suppressor
was used for the analysis of anions. All the anions were determined with isocratic elution at 1 mL/min of 12 mMNaOH elu-
ent. A Dionex SCS1 separation column with 3 mM methane sulfonic acid eluent was used for the analysis of cations. In order
to achieve quality control the samples were analyzed in duplicate. Routinely, blank sample was run after every 10 samples.
All the results reported in this paper were blank corrected.
4 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113

To characterize the particulate morphology high resolution scanning electron microscope (SEM) (FEI Quanta FEG 200)
was used. SEM stubs were prepared by cutting approximately 1 cm2 of filter paper and coating with gold using a sputter
coater. These SEM stubs then put into the SEM chamber for obtaining morphologies of individual particles.

2.4. CALPUFF modeling approach

In order to study the dynamics of air pollutants at the study site CALPUFF dispersion model was used to predict pollutant
concentrations during the first monitoring campaign. In the present study CALPUFF dispersion model version 6.4 was used to
ascertain the influence of line source and point source emissions on the PM10, SO2 and NO2 concentrations at NFR and FFR. A
domain of 30  30 km has been taken for modeling the dispersion pattern. A sampling grid of size 1  1 km was set over the
domain area. All the parameters required for CALMET, CALPUFF and CAL POST processors have been provided accordingly. In
CALMET processor, the surface air data and upper air data of study area for 1 year (2012) was used to setup the met grid. The
terrain elevation map and land use map of domain elevation have been downloaded from webgis.com. Meteorological data
was processed by CALMET processor for the period of 1 month between September and October 2012. The concentration and
deposition of pollutants has been predicted by CALPUFF processor. Emissions from line sources, point sources and area
sources for PM10, SO2 and NO2 were given as input for the processor. Model default values were used for dry and wet depo-
sition of PM10, SO2 and NO2.

2.4.1. Point sources

All the stacks inside the Hyundai Motors India Pvt. Ltd were considered as point sources. The stack monitoring was car-
ried out to find out the pollutant emissions. All the stacks were monitored for PM10, SO2 and NO2 using ECOTECH stack mon-
itoring kit.

2.4.2. Line sources

One-week traffic data for Chennai Bangalore national highway (NH4) was obtained from national highway authority of
India (NHAI). Percentage distribution of vehicle fleet at NH4 was calculated based on the vintage and type of fuel used
(CMDA, 2008). The emission loads from vehicles were calculated using the following equation.
X
E L  Veh  V  EF 1

where E is the mass emission of the pollutant per day, L is the length of the road in km, Veh is the number of vehicles per day,
V is the vintage of vehicles and EF is the mass emission factor in g/km (ARAI, 2007).

2.4.3. Area sources

All the minor roads inside the factory are collectively considered as area source. Traffic data for minor roads were
obtained from Hyundai Motors India Pvt. Ltd., Chennai.

3. Results and discussion

3.1. Air quality at NFR and FFR of the highway

The results indicated that the average PM10 concentration at both monitoring stations during January-February 2013 was
higher than that of September-October 2012. During first campaign the PM10 concentration at NFR and FFR were 60
124.2 lg/m3 (mean = 86.69 lg/m3, standard deviation = 18.56 lg/m3) and 40.23103.14 lg/m3 (mean = 73.16 lg/m3, stan-
dard deviation = 16.21 lg/m3), respectively where as in second campaign the PM10 concentration at NFR and FFR were
63.95141.36 lg/m3 (mean = 89.44 lg/m3, standard deviation = 18.69 lg/m3) and 46.98108.76 lg/m3 (mean = 81.91 lg/
m3, standard deviation = 16.42 lg/m3), respectively. Time series of PM10 at two monitoring stations were presented in
Fig. 2. In both the monitoring campaigns, PM10 concentration near to the highway is higher than the concentration away
from the highway. It was found that particle concentration was highly influenced by distance from the highway, with a large
spike in PM concentration next to the roadway that decays to lower levels at far field region. This indicates that the major
source of particles is emissions from vehicles. PM is both directly emitted, via combustion and mechanical wear, and formed
via secondary reactions of direct emissions, whereby organic and inorganic vapors undergo gas-to-particle conversion in the
atmosphere. In addition, re-suspension of particles due to strong traffic wake at NFR was also responsible for high PM con-
centration. At FFR the influence of traffic wake on dispersion of PM10 get diminished. Hence, we noticed less PM concentra-
tions at FFR. It was found that in most of the previous PM measurements showed sharp drop with distance away from
roadway (Kinney et al., 2011; Levy et al., 2003). Kinney et al. (2011) observed that the average PM concentrations near
the roadside were more than four times higher than average concentrations at 30 m away, and over six times higher than
those at 100 m away from the road.
During first monitoring period at NFR about 20% PM values were exceeding the national ambient air quality standards
(100 lg/m3) during second monitoring period it was about 23%. At FFR 10% and 12% PM values were exceeding NAAQS during
first and second monitoring campaign, respectively. Most of the time the PM concentrations at national highway (100%) and
 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113 5

NFR-S1 FFR-S1
PM10 NAAQS PM10 NAAQS
140 120

concentraon ( g/m3)
120 100
concentraon ( g/m3)

100 80
80 60
60
40
40
20
20
0
0

1-10-12
3-10-12
5-10-12
7-10-12
11-09-12
13-09-12
17-09-12
19-09-12
21-09-12
23-09-12
25-09-12
27-09-12
29-09-12

15-10-12
17-10-12
1-10-12

3-10-12

5-10-12

7-10-12
11-09-12

13-09-12

17-09-12

19-09-12

21-09-12

23-09-12

25-09-12

27-09-12

29-09-12

15-10-12

17-10-12
Days Days

NFR-S2 FFR-S2
160 120 PM10 NAAQS
PM10 NAAQS
140
concentraon ( g/m3)

100

concentraon ( g/m3)
120
100 80
80 60
60
40
40
20 20
0 0
18-01-2013

20-01-2013

22-01-2013

24-01-2013

28-01-2013

30-01-2013

01-02-2013

03-02-2013

05-02-2013

07-02-2013

09-02-2013

11-02-2013

13-02-2013

18-01-13

20-01-13

22-01-13

24-01-13

28-01-13

30-01-13

01-02-13

03-02-13

05-02-13

07-02-13

09-02-13

11-02-13

13-02-13
Days Days

Fig. 2. Time series of 24 h average PM10 measured at NFR and FFR for two seasons.

away from the highway (97%) are exceeding world health organization standards (PM10 = 50 lg/m3). The real time measure-
ment results indicated that the hourly maximum PM10 (169.44 lg/m3), PM2.5 (103.16 lg/m3) and PM1 (87 lg/m3) concentra-
tions were observed during morning peak hour traffic (around 9am) and minimum PM concentrations were found during
midnight (PM10 = 6.6 lg/m3, PM2.5 = 4.74 lg/m3, and PM1 = 3.1 lg/m3). The 24 h average PM10, PM2.5 and PM1 concentrations
at NFR were presented in Fig. 3. During the weekend the average concentration observed was (PM10 = 40 lg/m3,
PM2.5 = 25.48 lg/m3, PM1 = 20.44 lg/m3) lesser than the concentrations obtained during week days (PM10 = 50.49 lg/m3,
PM2.5 = 40.08 lg/m3, PM1 = 24.32 lg/m3). The PM10 comprises of about 70% of PM2.5 (80% of PM2.5 comprises of PM1) and
55% of PM1.
Similar to PM concentrations, the SO2 and NO2 concentrations were significantly high at NFR when compared to FFR. The
24 h average SO2 and NO2 concentrations measured at NFR and FFR were presented in Fig. 4. The SO2 concentration at NFR
was in the range of 8.923.5 lg/m3 and at FFR it varied from 5.88 to 24.36 lg/m3. The NO2 concentrations at NFR were varied
between 4.17 and 30.21 lg/m3 and at FFR it ranged between 9.25 and 30.21 lg/m3. The measured 24 h average NO2 and SO2
values were well below the NAAQS values (SO2 = 80 lg/m3 NO2 = 80 lg/m3).
The 24 h average meteorological parameters measured at the study site were presented in Table 1. The wind rose data
indicates that the air mass flows predominantly from south-west (SW) direction (Fig. 5). It was observed that the pollutant

90
PM10 PM2.5 PM1
80
concentraon g/m3

70

60

50

40

30

20

10

0
27-9-12

28-9-12

29-9-12

30-9-12

1-10-12

2-10-12

3-10-12

4-10-12

5-10-12

6-10-12

7-10-12

8-10-12

9-10-12

10-10-12

11-10-12

12-10-12

13-10-12

14-10-12

15-10-12

16-10-12

17-10-12

18-10-12

19-10-12

20-10-12

Date

Fig. 3. Time series of 24-h average PM10, PM2.5 and PM1 concentrations at NFR.
6 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113

25 30
SO2 near SO2 far SO2 near SO2 far

concentraon ( g/m3)
S1 S2
25

concentraon ( g/m3)
20
20
15
15
10
10
5 5

0 0

18-01-13
19-01-13
20-01-13
21-01-13
22-01-13
23-01-13
24-01-13
25-01-13
28-01-13
29-01-13
30-01-13
31-01-13

10-02-13
11-02-13
12-02-13
13-02-13
1-02-13
2-02-13
3-02-13
4-02-13
5-02-13
6-02-13
7-02-13
8-02-13
9-02-13
11-09-12
13-09-12
17-09-12
19-09-12
21-09-12
23-09-12
25-09-12
27-09-12
29-09-12

15-10-12
17-10-12
1-10-12
3-10-12
5-10-12
7-10-12
Days Days
35 35
NO2 near NO2 far NO2 near NO2 far
30 S1 30 S2
concentraon ( g/m3)

concentraon ( g/m3)
25 25

20 20
15 15
10 10
5 5
0 0
11-09-12
13-09-12
17-09-12
19-09-12
21-09-12
23-09-12
25-09-12
27-09-12
29-09-12

15-10-12
17-10-12
1-10-12
3-10-12
5-10-12
7-10-12

18-01-13
19-01-13
20-01-13
21-01-13
22-01-13
23-01-13
24-01-13
25-01-13
28-01-13
29-01-13
30-01-13
31-01-13

10-02-13
11-02-13
12-02-13
13-02-13
1-02-13
2-02-13
3-02-13
4-02-13
5-02-13
6-02-13
7-02-13
8-02-13
9-02-13
Days Days

Fig. 4. Time series of 24-h average SO2 and NO2 measured at NFR and FFR for two seasons.

Table 1
Summary of the 24 h average meteorological parameters.

Statistics Temperature (C) Wind speed (m/s) Relative humidity (%) Precipitation (mm)
Maximum 29.15 8.15 93.30 0.58 number of rainy days = 9/30
Minimum 26.97 2.87 64.94
Average 27.90 5.22 74.71
Standard deviation 1.2 1.2 6.34

Fig. 5. Wind rose diagram.

concentrations were inversely proportional to wind speed (i.e. pollutant concentration increases with decrease in wind
speed). However, on some occasions, high pollutant concentrations were also observed with increasing wind speed. As
mentioned earlier, the predominant wind direction at the study area is SW, when wind direction was SW and wind
speed was low, high pollutant concentrations was observed at study site. Srimuruganandam and Nagendra (2011a) also
observed an inverse relation of PM with wind speed at SP road, Chennai. It was found that high pollutant concentrations
(PM10 = 99 lg/m3, SO2 = 18.8 lg/m3, NO2 = 19.97 lg/m3) corresponding to minimum temperature (26.97 C) and minimum
values (PM10 = 78 lg/m3, SO2 = 1.55 lg/m3, NO2 = 10.31 lg/m3) corresponding to maximum temperature (29.15 C). The
 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113 7

results indicated that increase in percentage humidity, increases the PM levels. Since the Sriperumbadur is close to Bay of
Bengal the ambient particles contain moisture even on relatively dry days. Under humid condition, the particles absorb more
moisture and increases in size and volume. This situation leading to increases in settling velocities of the most particles with
leads to dry deposition. However, SO2 and NO2 were not showing any significant relationship with humidity.

3.2. Particulate matter chemical characteristics

The PM elemental analysis showed that the elements such as Na, Al, Ca, Mn, K, Zn, and Fe had the highest contributions to
PM at the study area (Fig. 6). The Na levels were considerably higher than the other elements. Since the study area is located
near Bay of Bengal, sea spray was considered a contributor to Na. In addition, geological materials can also contribute sig-
nificantly to Na. It was observed that the concentrations of crustal elements (Al, Fe, K, Mg and Ca) were much higher than
those of toxic elements for both the monitoring stations (NFR and FFR). The Al can also emit from the Aluminum foundry
present in the Hyundai factory. However, its concentration was found to be higher at NFR when compared to FFR, confirming
the contribution from re-suspension of soil dust due to vehicular movement. The elements Cu, Pb, and Li were found to be at
lower concentrations, but Ba and Zn concentrations were found to be significant. Ba is added to lubricating oil to prevent
smoke and engine abrasion of diesel vehicles. Zn is emitted wear and tear of tires. It was observed that most of the elements
were found higher at NFR when compared to FFR. So the common sources recognized to these elements at this site is
re-suspension of road dust due to vehicular activities like wear and tear of tires, oil burning, abrasion of mechanical parts
of vehicles, and oil lubricants.
Particulate ionic concentrations in both monitoring stations were presented in Fig. 7. Among the water soluble ions, NO
3



was found to be the largest component of PM. Other abundant components included NH+3, SO2
4 , Cl , and NO2 ions. The

14
12 NFR
Concentraon g/m3

10
8
6
4
2
0
Al B Ba Ca Cr Cu Fe K Mg Mn Na Ni Pb Sr Zn
elements

12

10 FFR
concentraon g/m3

0
Al B Ba Ca Cr Cu Fe K Mg Mn Na Ni Pb Sr Zn
elements

Fig. 6. Elemental composition of PM10 at NFR and FFR.

8 9
7 NFR 8 FFR
concentraon g/m3
concentraon g/m3

6 7
5 6
5
4
4
3
3
2 2
1 1
0 0
K

F-
Ca

Cl-
Na

Br-
Mg

SO4

PO4
NH4

NO3
NO2-
K

F-
Ca

Cl-
Na

Br-
Mg

SO4

PO4
NH4

NO3
NO2-

ions ions

Fig. 7. Ionic composition of PM10 at NFR and FFR.

8 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113

presence of SO2
4 and NO3 concentrations proves the involvement of the secondary particulate formations and photochem-


ical reaction. At FFR, SO2

4 and NO3 were found to be 18 and 35% higher than NFR indicating the conversions of SO2 and NO2
to secondary particles. The SO2 and NO2 concentrations were 23 and 21% lower at FFR when compared to NFR. SO2 is
depleted mostly by dry deposition, and by conversion to SO2
2

4 . The proportion of SO4 aerosol in the atmosphere depends
upon meteorological conditions and local concentrations of NH3 (Hassan et al., 2013). NO
3 exists mostly in the form of
ammonium nitrate (NH4NO3). NH4NO3 is unstable under atmospheric conditions, existing in reversible phase equilibria with
gaseous precursors, nitric acid and ammonia. The dissociation constant of the equilibria depends on temperature and relative
humidity (Yeatman et al., 2001; Lin et al., 2010). Soil is considered to be the main source of Mg2+, Ca2+ and K+ at the study
site. The other source for Mg2+, Ca2+ and K+ is the sea spray (Kim et al., 2006).
In general, our analysis and results illustrate that the concentrations of PM10 are quite high for an Indian cities and much
higher than the air pollution standards set by CPCB. The chemical constituents in PM10 of our measurement data is compared
those reported previously from other urban, industrial, background & rural sites around the world (Table 2). Toxic trace ele-
ments have been investigated intensively for many years because of their adverse health effects. The elemental profile at
urban sites was dominated by terrestrial elements (i.e., Fe, Ca, Al, Mg) and sea salt (Na, Cl). Natural contributions and
long-range transport processes were found to play a key role on the air pollution in rural and urban background areas. At
industrial site, elemental composition showed that the PM10 was dominated by terrestrial elements and trace elements
emitted by the industries (Table 2). But our study indicates that the major source of trace elements was traffic emissions.
To the best of our knowledge, there are no studies reported about the chemical composition PM10 at near field and far field
region of the national highway under tropical climate.

3.3. Particulate morphological characteristics

Fig. 8 represents the scanning electron microscopy (SEM) images of the PM collected at the study site. Qualitative assess-
ment of the images shows that most of the images are of small chains like spherical particles. Spherical particles (A) repre-
sents the soot particles which are indicative of combustion or other high temperature processes. Soot particles are
characterized by unique morphology ranged from linear to more branched aggregates of small spherical particles. These
are carbonaceous material emitted from the incomplete combustion of oil, gasoline, diesel, fuel oil and coal. The soot aggre-
gates obtained in this study were analogous to soot particles emitted from the vehicular emissions. These are the typical par-
ticle type originating from vehicle exhaust emissions (Chithra and Nagendra, 2013). Typical diesel particles are agglomerates
consisting of mainly spherical primary particles of about 1540 nm in diameter. The irregular shaped particle (B) of size
about 5 lm indicates the mineral particles which are of crustal origin. Mineral particles are characterized by the dominance
of Si and Al with minor concentrations of Ca, Fe, Na, Mg (Jung et al., 2010; Lu et al., 2006; Okada et al., 2005; Shao et al., 2008;
Shi et al., 2005, 2003; Iwasaka et al., 2003; Srimuruganandam and Nagendra, 2011b). They are originated mainly from crustal
source, but they are also emitted from construction activities, and resuspension of road dust. The morphological structure of
these particles varied from irregular to regular shapes. The microanalysis study of fine PM at Chennai city also showed the
presence of mineral particles with dominance of Si, Al, Ca, Na and Fe. Porous particle (C) represents pollen and collection of
one or more crystalline structure (D) represents salts or marine aerosol. Sea salt particles tend to be highest in coastal pol-
luted air. Since the study area is located near Bay of Bengal, sea spray was considered a contributor of marine aerosols. The
sea salt particles are also found to react with polluted air, which can lead to high surface ozone levels in turn affecting the air
quality at coastal cities (von Glasow, 2008).

3.4. Traffic characteristics

Vehicles were categorized under various groups viz. two wheelers (2W), three wheelers (3W), Light Commercial Vehicles
(LCV) and Heavy Commercial Vehicles (HCV). Fig. 9 shows the distribution of heterogeneous traffic composition and vehicle

Table 2
Details of stacks in the Hyundai factory and the emission rate.

Sl. no. Stack Pollutants Emission Emission Exit temperature Velocity Height Diameter
(mg/N m3) (g/s) (K) (m/s) (m) (m)
1. Aluminum foundry 1 PM10 6.82 0.04 323 8.63 15 0.93
SO2 0.00 0.00
NO2 4.77 0.03
2. Boiler PM10 47.14 0.11 503 9.74 58 1.96
SO2 55.27 0.13
NO2 18.68 0.05
3. Incinerator PM10 35.20 0.09 341 10.15 32 0.59
SO2 45.00 0.11
NO2 16.52 0.04
4. Aluminum foundry 2 PM10 8.55 0.02 321 8.6 15 1.4
SO2 0.00 0.00
NO2 5.72 0.01
 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113 9

Fig. 8. Scanning electron microscopy image of PM.

Fig. 9. Percentage distribution of heterogeneous traffic composition at NH4.

classification (based on vintage) at Chennai Bangalore national highway. Analysis of traffic flow on NH4 indicated that vehi-
cle density at the study site was different from typical urban traffic patterns. Traffic fleets were comprised mostly of heavy
duty vehicles. The data analysis showed that 4 wheeler diesel vehicles (62%) were dominating followed by 4 wheeler petrol
vehicles (22%). All other groups of vehicles were representing rest 16%. It was found that significant increase in 4 wheeler
diesel vehicles during week days when compared to weekend. In general, peak morning and evening traffic flows occurred
10 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113

between 911 am and 25 pm, respectively. But peak flow for individual category of vehicles was found to be different. For
cars and two wheelers the peak hour was between 9 and 11 am, for buses the peak hours was between 8 and 10 am where as
trucks showed very small fluctuation with respect to time. Total emissions of PM10, SO2 and NO2 from vehicles were found to
be 25 kg/day, 4.26 kg/day and 164 kg/day, respectively.

3.5. Stack emissions

The emissions from different stacks were given in Table 3. The stack monitoring results shows that maximum concentra-
tion of all the pollutants observed at stack 2 (PM10 = 47.14 lg/N m3, SO2 = 55.27 lg/N m3, and NO2 = 18.68 lg/N m3) and
minimum at stack 4 (PM10 = 8.55 lg/N m3, SO2 = 0 lg/N m3, and NO2 = 5.72 lg/N m3). The SO2 emissions from the boiler
and incinerator stacks were found predominant and the PM10 emissions were found maximum in aluminum foundry stack.

3.6. CALPUFF modeling

To understand the dispersion pattern of PM10, SO2 and NO2, CALPUFF model was used. The predicted PM10, SO2 and NO2
concentrations at NFR and FFR were found to be 32.31 1.56 and 31.35 1.27 lg/m3; 0.37 0.21 and 0.06 0.04 lg/m3; and
12.82 6.55 and 4.65 2.77 lg/m3, respectively (Table 4). The concentrations of pollutants were found to maximum when
wind direction is South West and wind speed is less than 2 m/s. The dispersion pattern of 24 h average PM10, SO2 and NO2
concentration over the domain were presented in Fig. 10. From the dispersion pattern, maximum of 3, 99, and 98% reduction

Table 3
Summary of major chemical constituents in PM10 reported in the past studies.

Study area Particle groups Source Reference

Mt. Qomolangma  Soot/tar ball  Fuels, biomass burning, incomplete fossil fuel, Cong et al. (2010)
(Himalayas)  Aluminosilicates combustion
 Calcium sulfate  Soil
 Ca/Mg CO3  Local crustal source
 Fe/Ti-rich particles  Soil
 Pb-rich particles  Soil
 Automobile, mining, waste incineration
Portugal  Biological particles  Pollen, bacteria, excrements of insects Slezakova et al. (2008)
(background)  Aluminum silicates  Soil
 Fe oxides and alloys  Automobile and traffic
 Sulfates  Sea
 Ti/Al rich  Soil
 Lead sulfate  Traffic
Pune, India  Aluminosilicates  Combustion process & crustal (soil)
 Carbonates (CaCO3)  Erosion of rocks
 Halite (NaCl)  Sea
 Sulfate  Biomass burning and fossil fuel
 Soot  Burning and combustion of fossil fuel
Agra, India  Aluminosilicates  Crustal Pipal et al. (2011)
 Ca rich (CaCO3)  Construction, agriculture, natural soil
 C-rich particles (soot and tar balls)  Gasoline and coal burning
Xinjiang, China  Mineral particles  Crustal source Talifu et al. (2015)
 Soot particles  Coal combustion;
 Fly ash  Vehicular exhaust
 Coal combustion
Chennai, India  Soot particles  Combustion Srimuruganandam and
 Mineral particles  Crustal origin Nagendra (2011a), Chithra and
 Salts  Marine aerosols Nagendra (2013), Present study

Table 4
Performance statistics of CALPUFF model.

Station Pollutants Predicted-P (lg/m3) Observed-O (lg/m3) RMSE sys RMSE unsys MBE R2 STDEV-P STDEV-O IA
Near Field PM10 32.31 76.37 7.35 17.29 8.52 0.519 1.56 7.63 0.73
NO2 12.83 13.39 8.62 5.85 12.92 0.531 0.21 4.11 0.98
SO2 0.37 13.29 3.99 4.33 2.33 0.457 6.55 5.90 0.20
Far field PM10 31.35 73.37 7.20 51.66 42.03 0.443 1.27 16.46 0.94
NO2 4.65 11.59 8.78 3.38 10.56 0.410 0.04 2.96 0.99
SO2 0.06 10.62 6.59 4.27 6.94 0.614 2.77 5.31 0.47
 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113 11

Fig. 10. Dispersion pattern of 24 h average PM10, SO2 and NO2 concentration over the domain.
12 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113

140 100
PM10 FFR PM10 NFR
120
measured (g/m3) 80

measured (g/m3)
100

80 60
R = 0.4431
60 40 R = 0.5196
40
20
20

0 0
29 31 33 35 37 39 32 34 36 38 40
predicted (g/m3) predicted (g/m3)
12 35
NO2 FFR NO2 NFR
30
10
Measured (g/m3)

Measured (g/m3)
25
8
20
6
R = 0.6149 15
R = 0.4574
4
10
2
5
0 0
0 5 10 15 20 25 30 0 10 20 30 40
predicted (g/m3) predicted (g/m3)

18 1.2
SO2 FFR SO2 NFR
16
Measured (g/m3)

1.0
Measured (g/m3)

R = 0.4102
14
0.8
12
0.6
10
R = 0.5316 0.4
8
6 0.2

4 0.0
0.0 0.1 0.1 0.2 0 5 10 15 20 25
predicted (g/m3) predicted (g/m3)

Fig. 11. Scatter plot of measured VS predicted pollutant concentrations.

of PM10, SO2 and NO2 concentrations were observed between NFR and FFR, respectively. The reduction in PM10 concentration
was found to be less because stack 1 (aluminum foundry) was located near to FFR, which emits considerable amount of PM.
All other stacks were located almost equally from both NFR and FFR. The percentage reduction in observed PM10 was 16%
which was not reflected in model predictions. In general, model underestimated PM10 concentrations mainly because of
the re-suspension of particles due vehicular movement (inside and outside factory), which was not taken into account by
the model. Further, the observed SO2 and NO2 concentrations were 23 and 21% lower at FFR when compared to NFR, which
is over estimated by the model. Model performance was evaluated on the basis of statistical indicators namely correlation
coefficient (R2), index of agreement (IA) and root mean square error (RMSE) (Table 4). Fig. 11 presents the scatter plot of
measured and predicted PM10, SO2 and NO2 concentrations at the study area. It can be seen that the results showed a mod-
erate correlation between the predicted and measured PM10 (R2 = 0.440.52), SO2 (R2 = 0.410.51) and NO2 (R2 = 0.450.61)
concentrations. In general, the model predictions were comparable with the observed data.

4. Conclusions

The physical and chemical characteristics of PM at NFR and FFR of a highway passing through an industrial complex were
investigated. Results showed high PM10, SO2 and NO2 concentrations at NFR than FFR from the highway. The PM10 concen-
trations at NFR and FFR were exceeding NAAQS by 2023% and 1012%, respectively. The chemical characterization of PM10
showed the abundance of major elements such as Na (30%), Ca (12%), and B (12%) at NFR and Na (32%), Ca(14%), Fe(18%), etc.


at FFR. At FFR, SO2

4 and NO3 were found to be 18 and 35% higher than NFR indicating the conversions of SO2 and NO2 to
 S.M. Shiva Nagendra et al. / Transportation Research Part D 48 (2016) 113 13

secondary particles. The measured SO2 and NO2 concentrations were 23 and 21% lower at FFR when compared to NFR. The
simulation made by the CALPUFF model showed reasonable accuracy for PM10 (R2 = 0.440.52), SO2 (R2 = 0.410.51) and NO2
(R2 = 0.450.61) concentrations.
The present study, indicate that emissions from heavy duty vehicles contributes significantly to local air pollution prob-
lem in the industrial complex. The chemical composition of PM at industrial area can provide important information for
future assessment of the health risk to the local population and prioritization of control measures.

Acknowledgement

We wish to thank the Hyundai Motors India Private Limited and Head, Sophisticated Analytical Instruments Facility
(SAIF), Indian Institute of Technology Madras (IITM), Chennai, India for their support to this study.

References

Atkinson, R., 1990. Gas phase tropospheric chemistry of organic compounds: a review. Atmos. Environ. 24, 141.
Automotive Research Association of India (ARAI), 2007. Draft Report on Emission Factor development for Indian Vehicles as a Part of Ambient Air Quality
Monitoring and Emission Source Apportionment Studies.
Central Pollution Control Board (CPCB), 2009a. National Ambient Air Quality Monitoring Series. Central Pollution Control Board, New Delhi.
Central Pollution Control Board (CPCB), 2009b. Comprehensive Environmental Assessment of Industrial Clusters. NewDelhi.
Central Pollution Control Board (CPCB), 2010. Air Quality Monitoring, Emission Inventory and Source Apportionment Study for Indian Cities. New Delhi.
Central Pollution Control Board (CPCB), 2011. Guidelines for the Measurement of Ambient Air PollutantsVolume-I. New Delhi.
Chennai Metropolitan Development Authority (CMDA). 2008. Second Master Plan, Chennai.
Chithra, V.S., Nagendra, S.M.S., 2013. Chemical and morphological characteristics of indoor and outdoor particulate matter in an urban environment. Atmos.
Environ. 77, 579587.
Cong, Z., Kang, S., Dong, S., Liu, X., Qin, D., 2010. Elemental and individual particle analysis of atmospheric aerosols from high Himalayas. Environ. Monit.
Assess. 160 (14), 323335.
Hassan, S.K., ElAbssawy, A.A., Khoder, M.I., 2013. Characteristics of gasphase nitric acid and ammoniumnitratesulfate aerosol, and their gasphase
precursors in a suburban area in Cairo, Egypt. Atmos. Pollut. Res. 4, 117129.
Iwasaka, Y., Shi, G.Y., Yamada, M., Matsuki, A., Trochkine, D., Kim, Y.S., Zhang, D., Nagatani, T., Shibata, T., Nagatani, M., Nakata, H., 2003. Importance of dust
particles in the free troposphere over the Taklamakan Desert: electron microscopic experiments of particles collected with a balloonborne particle
impactor at Dunhuang, China. J. Geophys. Res.: Atmos. 108 (D23).
Jung, H.J., Malek, M.A., Ryu, J., Kim, B., Song, Y.C., Kim, H., Ro, C.U., 2010. Speciation of individual mineral particles of micrometer size by the combined use of
attenuated total reflectance-Fourier transform-infrared imaging and quantitative energy-dispersive electron probe X-ray microanalysis techniques.
Anal. Chem. 82 (14), 61936202.
Kim, K.H., Mishra, V.K., Kang, C.H., Choi, K.C., Kim, Y.J., Kim, D.S., 2006. The ionic compositions of fine and coarse particle fractions in the two urban areas of
Korea. J. Environ. Manage. 78, 170182.
Kinney, P.L., Gichuru, M.G., Volavka-Close, N., Ngo, N., Ndiba, P.K., Law, A., Gachanja, A., Gaita, S.M., Chillrud, S.N., Sclar, E., 2011. Traffic impacts on PM2.5 air
quality in Nairobi, Kenya. Environ. Sci. Policy 14, 369378.
Levy, J.I., Bennett, D.H., Melly, S.J., Spengler, J.D., 2003. Influence of traffic patterns on particulate matter and polycyclic aromatic hydrocarbon
concentrations in Roxbury, Massachusetts. J. Eposure Sci. Environ. Epidemiol. 13, 364371.
Lin, Y.C., Cheng, M.T., Lin, W.H., Lan, Y., Tsuang, B., 2010. Causes of the elevated nitrate aerosol levels during episodic days in Taichung urban area, Taiwan.
Atmos. Environ. 44, 16321640.
Lu, S.L., Shao, L.Y., Wu, M.H., Jiao, Z., 2006. Mineralogical characterization of airborne individual particulates in Beijing PM10. J. Environ. Sci. 18, 9095.
Ntziachristos, L., Ning, Z., Geller, M.D., Sioutas, C., 2007. Particle concentration and characteristics near a major freeway with heavy-duty diesel traffic.
Environ. Sci. Technol. 41, 22232230.
Okada, K., Qin, Y., Kai, K., 2005. Elemental composition and mixing properties of atmospheric mineral particles collected in Hohhot, China. Atmos. Res. 73,
4567.
Pipal, A.S., Kulshrestha, A., Taneja, A., 2011. Characterization and morphological analysis of airborne PM 2.5 and PM 10 in Agra located in north central India.
Atmos. Environt. 45 (21), 36213630.
Shao, L.Y., Li, W.J., Xiao, Z.H., Sun, Z.Q., 2008. The mineralogy and possible sources of spring dust particles over Beijing. Adv. Atmos. Sci. 25, 395403.
Shi, Z.B., Shao, L.Y., Jones, T.P., Lu, S.L., 2005. Microscopy and mineralogy of airborne particles collected during severe dust storm episodes in Beijing, China. J.
Geophys. Res. 110, 1303-1303.
Shi, Z., Shao, L., Jones, T.P., Whittaker, A.G., Lu, S., Berube, K.A., He, T., Richards, R.J., 2003. Characterization of airborne individual particles collected in an
urban area, a satellite city and a clean air area in Beijing, 2001. Atmos. Environ. 37 (29), 40974108.
Slezakova, K., Pires, J.C.M., Pereira, M.C., Martins, F.G., Alvim-Ferraz, M.C., 2008. Influence of traffic emissions on the composition of atmospheric particles of
different sizesPart 2: SEMEDS characterization. J. Atmos. Chem. 60 (3), 221236.
Srimuruganandam, B., Nagendra, S.M.S., 2011a. Characteristics of particulate matter and heterogeneous traffic in the urban area of India. Atmos. Environ. 45,
30913102.
Srimuruganandam, B., Nagendra, S.M.S., 2011b. Chemical characterization of PM 10 and PM 2.5 mass concentrations emitted by heterogeneous traffic. Sci.
Total Environ. 409 (17), 31443157.
Talifu, D., Wuji, A., Tursun, Y., Kang, H., Hu, Y., Guo, Y., Shao, L., 2015. Micro-Morphological Characteristics and Size Distribution of PM 2.5 in the Kuitun-
Dushanzi Region of Xinjiang, China. Aerosol Air Qual. Res. 15 (6S), 22582269.
USEPA, 1998. A Comparison of CALPUFF Modeling Results to Two Tracer Field Experiments <http://www.epa.gov/scram001/7thconf/calpuff/tracer.pdf>.
USEPA, 1999a. Compendium of Methods for the Determination of Inorganic Compounds in Ambient Air, Compendium Method IO-3.1: Selection, Preparation
and Extraction of Filter Material.
USEPA, 1999b. Compendium of Methods for the Determination of Inorganic Compounds in Ambient Air, Compendium Method IO-3.5: Determination of
Metals in Ambient Particulate Matter using Inductively Coupled Plasma/Mass Spectrometry (ICP/MS).
USEPA, 2000. Requirements for preparation, adoption, and submittal of state implementation plans (Guideline on Air Quality Models); proposed rule. Fed.
Reg. 65, 2150521546.
USEPA, 2008. Clarification of Regulatory Status of CALPUFF for Near-Field Applications. U.S. Environmental Protection Agency, Research Triangle Park, NC.
von Glasow, R., 2008. Pollution meets sea salt. Nat. Geosci. 1, 292.
Yeatman, S.G., Spokes, L.J., Jickells, T.D., 2001. Comparisons of coarse-mode aerosol nitrate and ammonium at two polluted coastal sites. Atmos. Environ. 35,
13211335.