Science of the Total Environment 408 (2010) 18001811

Contents lists available at ScienceDirect

Science of the Total Environment

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v

Estimation of automobile emissions and control strategies in India

K.S. Nesamani
Institute of Transportation Studies, University of California Irvine, CA 92612, United States

a r t i c l e

i n f o

Article history:
Received 9 April 2009
Received in revised form 11 January 2010
Accepted 18 January 2010
Available online 11 February 2010
Keywords:
Emission inventory
IVE model
GPS
Global warming potential
Emission-reduction strategies
Driving patterns
Chennai

a b s t r a c t
Rapid, but unplanned urban development and the consequent urban sprawl coupled with economic growth
have aggravated auto dependency in India over the last two decades. This has resulted in congestion and
pollution in cities. The central and state governments have taken many ameliorative measures to reduce
vehicular emissions. However, evolution of scientic methods for emission inventory is crucial. Therefore, an
attempt has been made to estimate the emissions (running and start) from on-road vehicles in Chennai
using IVE model in this paper. GPS was used to collect driving patterns.
The estimated emissions from motor vehicles in Chennai in 2005 were 431, 119, 46, 7, 4575, 29, and 0.41 tons/
days respectively for CO, VOC, NOx, PM, CO2, CH4 and N2O. It is observed from the results that air quality in
Chennai has degraded. The estimation revealed that two and three-wheelers emitted about 64% of the total CO
emissions and heavy-duty vehicles accounted for more than 60% and 36% of the NOx and PM emissions
respectively. About 19% of total emissions were that of start emissions. It is also estimated that on-road
transport contributes about 6637 tons/day CO2 equivalent in Chennai. This paper has further examined various
mitigation options to reduce vehicular emissions. The study has concluded that advanced vehicular technology
and augmentation of public transit would have signicant impact on reducing vehicular emissions.
2010 Elsevier B.V. All rights reserved.

1. Introduction
Air quality in developing countries like India has reached an
alarmingly low level. Most cities have exceeded the National Ambient
Air Quality (NAAQ) standards. Particulate matter (PM) is a major
concern in Indian cities and 60 out of 62 metropolitan cities have
exceeded World Health Organization (WHO) standards (24-h ambient
air quality standards). Air pollution has become one of the leading
causes of death in India. Pachauri and Sridharan (1998) estimated that
the pollution load increased from 0.15 million tons in 1947 to
10.3 million tons in 1997 from the transport sector alone. CO claimed
the largest share (43%) in the total, followed by NOx (30%), HC (20%),
SPM (5%), and SO2 (2%). Transport's emissions have increased at the
faster rate than any other sector. Twenty percent of poorly maintained
vehicles produce about 60% of vehicular pollution in India (Pundir,
2000). Main causes for the shocking increase in vehicular emissions
have been the exponential growth in the number of motor vehicles;
inadequate public transport and inept management; haphazard urban
development; congestion; obsolete vehicular technology; poor fuel
quality; laxity in trafc enforcement; and an increase in freight moved
over roads (The World Bank, 2005; Badami, 2005; Pucher et al., 2007).
Fig. 1 shows that, as the number of cities grows, vehicle population
and travel demand increases, driving up the amount of pollution as
well. However, road length grows slowly. This signicant disparity

E-mail address: [email protected].

0048-9697/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.scitotenv.2010.01.026

between demand and supply of transport infrastructure has caused

severe congestion in all major Indian cities during peak hours. It has
not only increased delays but also air pollution.
In recent years, the increasing contribution of the transport sector to
greenhouse gas (GHG) emissions resulting from the use of fossil fuels,
and their effects on global warming and climate change have been a
major concern. According to the world Bank, in developing countries the
share of GHG emissions are low, but the energy consumption within the
transport sector is growing much faster than other sectors (Karkezi
et al., 2003). It is estimated that there are about 900 million vehicles
(excluding two-wheelers) worldwide that emit more than 26% of GHG
emissions (Lee Chapman, 2007). In India, CO2 emissions from the
consumption of fossil fuels have increased from 293 MMT (million
metric tons) to 1293MMT between 1980 and 2006 (EIA, 2006). Road
transport accounts for nearly 35% of commercial fuel sold in India. ADB
projects the CO2 emission from on-road transport in India will increase
about 600% in next three decades (20052035) (ADB, 2006).
It is crucial to accurately estimate emissions. There are many emission
models such as LEAP (Long-range Energy Alternative Planning), VAPIS
(Vehicle Air Pollution Information System) and Spreadsheet models to
estimate vehicular emissions (Bose and Nesamani 2001), (Kokaz et al.,
2001), (Gurjar et al., 2004). However, these models don't capture
variations in local environments such as driving behaviors and trafc
conditions. Further, these models are unable to assess the impact of
trafc-management schemes such as roundabouts, signal coordination,
and road widening. These models also assume that lower vehicle speeds
increase emissions. However, researchers have established that higher

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

1801

Fig. 1. Relative increase in the infrastructure, vehicle population, energy consumption, and travel demand during the last ve decades (19512000) in India.

speeds can also increase or decrease emission if conditions are not steady
(Anh et al., 2002). In addition, emission rates in these models are
inadequate and fail to represent the latest vehicle mix and operating
conditions.
To control vehicular emissions central, state and local governments
have implemented many policy measures such as improving vehicular
technology, revising trafc-management schemes, implementing
stricter emission controls, introducing cleaner fuels, and promoting
alternative fuels (CNG and LPG). The effectiveness of such efforts has
not been evaluated. Consequently, the objectives of this paper are
twofold: (i) to develop a methodology to estimate vehicular emissions
and (ii) to evaluate existing strategies to control vehicular emissions
and recommend alternate strategies. The objectives could be achieved
with credible database and sophisticated analytical framework.
Research studies (Joumard et al., 1995; Sjodin and Lenner, 1995)
indicate that driving patterns signicantly affect vehicular emissions.
Therefore, in this study, global positioning systems (GPS) were used to
collect driving patterns. Chennai City was selected for the case study.
2. Transport characteristics of Chennai City
Chennai City is one of four metropolitan cities in India. It is located
on the southeastern coast and as reported in the 2001 census a
population of 4.34 million people spread over 176 km2. The population
density was 24,681/km2. Its economic base has ranged from utility
service to automobiles. The city primarily depends on road transport
and suburban rail services for intra-city commuting. The road network
is in the form of radial pattern. N.H. 5 in the north, N.H. 45 in the
southwest, and N.H.4 and N.H. 205 in the west radiate outwards from
the Central Business District. Inner Ring Road and NH Bypass are other
major corridors. The city has less than 10% of the space for roads and
about 28% of road space is used for parking (WSA, 2008). Two-lane
and four-lane roads account for 57% and 37% respectively.
Each household completes an average of about six trips per day,
56% for work and education, 23% for shopping, and the balance for
other trips (Srinivasan and Rogers, 2005). The pollutants per day from
road transport in 2002 were 177 metric tons of carbon monoxide
(CO), 96 metric tons of hydrocarbons (HC), 27 metric tons of oxides of
nitrogen (NOx) and 7 metric tons per day of particulate matter (PM)
(GOI, 2002). Gasoline and diesel are the most common fuels in
Chennai with a disproportionately higher share of diesel vehicles. On
average, there were about 385 vehicles per 1000 people in 2002. This
was mainly due to rapid growth in the share of two-wheeled vehicles.
An average of 425 new vehicles are added every day (CMDA, 2007).
Table 1 describes the city's characteristics.

3. Methodology
The accuracy of emission inventory depends upon the availability
of reliable local data such as the number of on-road vehicles, eet
characteristics, vehicle utilization, and driving patterns. These data are
not readily available for Chennai City. Therefore, a primary survey was
necessary. Surveys of driving patterns and a questionnaire were
conducted to collect data. It is not feasible to collect vehicle activity
data in all areas of the city. Therefore, the city was broadly classied as
Central Business District, Residential, and Industrial for the purposes
of the study. Representative roads were selected.
The proposed methodology has three steps as illustrated in Fig. 2.
Estimate the travel demand for different vehicle classes.
Assess the eet composition of on-road vehicles.
Use the identied emission model to estimate emissions.
It is essential to know the number of on-road vehicles in order to
estimate the travel demand. Hence in this study vehicle population was
regressed based on the GDP growth rate as shown in Eq. (1). The growth
in real per capita income in Chennai has shown strong inuence on the
ownership of passenger vehicles and intermediate public transport. The
registration of buses and goods vehicles is strongly inuenced by the
Table 1
Transport characteristics of the city.
City population (2001)a
Areab
Population density
Number of registered vehicles (2007)b
Modal splitc

Fare (Rs.)
Annual fuel consumption (2001)d
Speed limit
Road length (2007)b
Average trip lengthc
Trip rate (2008)c
Number of motor vehicle accidents (2007)b
a
b
c
d

4.34 million
174 km2
24,231 persons per km2
2.65 million
Private vehicles 30%
IPT 8%
Public transport 31%
NMT 31%
Bus 2.00 (USD 0.04)
Rail 4.00 (USD 0.08)
Diesel 298 TMT
Petrol 152 TMT
Cars and two-wheelers 40 km/h
Three-wheelers and buses 25 km/h
618.3 km
11 km
1.5 (all modes)
1.02 (motorized)
7570

Census of India, (http://censusindia.gov.in/).

Statistical Handbook 2008, Government of Tamil Nadu.
WSA, 2008.
Sengupta, 2005.

1802

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

Table 2
Global warming potential of different pollutants.
Pollutant

GWP 100 year

Carbon dioxide (CO2)

Methane (CH4)
Nitrous oxide (N2O)
Carbon monoxide (CO)
Non-methane volatile organic compound (NMVOC)
Nitrogen oxide (NOx)a

1
25
298
1.9
3.4
NA

Fig. 2. Methodology to estimate vehicular emissions.

increase in the net state domestic product of Tamil Nadu. Vehicle

population data for different categories of vehicles were adjusted using
the average lifespan in the absence of any data regarding the number of
vehicles retiring from a road. Lifespan of different vehicles were dened
as follows:

Two-wheelers 15 years
Three-wheelers 12 years
Automobiles, taxis, and SUVs 15 years
Buses 10 years
Goods vehicles 15 years

for the period of 1990 to 2050. It can estimate emissions from more
than 700 different types of technologies with different combinations
of fuel and after-treatment technologies.
It accounts separately for start emissions and running emissions. It
also further considers changes in emission rates over time due to eet
turnover, diurnal emissions, hot-soak emissions, running losses, and
refueling emissions. Base emission rates depend on vehicle technology,
air/fuel ratios, engine sizes, and fuel types. The base emission rate data
are collected from different locations such as U.S., Europe, China, India,
and Thailand. It uses Vehicle Specic Power (VSP) and engine stress to
capture the impact of driving behavior more accurately. VSP is estimated
using speed, acceleration, and grade and it is shown in Eq. (3).
The IVE model estimates emission rates using Eq. (4) by adjusting
for different correction factors. The general inputs to the model are
eet characteristics, vehicle activity, driving patterns, fuel quality, and
temperature based on local conditions (IVE, 2008, Davis et al., 2005).
VSP = v1:1a + 9:81a tansingrade + 0:132 + 0:000302v

where, VSP Vehicle Specic Power.

grade = ht = 0 ht = 1 = vt = 1 to 0

[Mittal and Sharma, 2000; Bose and Nesamani, 2001].

The estimated on-road population was multiplied by vehicle
utilization to compute the travel demand as shown in Eq. (2).
V = a + bXi

Pit = Vit Uit

v
a
h

velocity (m/s)
acceleration (m/s2)
altitude (m).

Q t  = Bt K Baset K Tmpt K Hmdt K IMt 

K Fuelt K Alt tK CntrytK dt 
where,

where,
P
V
U
i
t
X

GWP for NOx is highly uncertain and hence it is not included.

annual vehicle km traveled

number of vehicles on road
annual vehicle utilization (km)
vehicle mode (car, TW, 3-wh, bus and trucks)
calendar year
denotes the real per capita income corresponding to 2wheelers, cars/jeeps and 3-wheelers. But, for buses and trucks,
X denotes the net state domestic product at 1980/81 prices.

On-road eet characteristics are one of the important factors to

estimate emissions. In this study, eet characteristics were rened
based on (i) the time series sales data of different technologies within
a given mode; and (ii) the previous data collection by Central Road
Research Institute (CRRI, 2002) and Pallavan Transport Consultancy
Services (PTCS, 1995). Fuel quality, temperature, and humidity were
collected from the secondary survey.
The International Vehicle Emission (IVE) model is used to estimate
emission inventory in this study (IVE, 2008). IVE model was jointly
developed by International Sustainable Systems Research Center and
the University of California at Riverside through funding from the
United States Environmental Protection Agency. This model calculates
vehicular emissions at macro, meso, and micro scales. The IVE model
can estimate emission inventory for local, global and toxic pollutants

Q[t]
adjusted emission rate for each technology
B[t]
base emission rate in for each technology
K(Base)[t]
adjustment to the base emission rate
K(Tmp)[t]
temperature correction factor
K(Hmd)[t]
humidity correction factor
K(IM)[t] inspection/maintenance correction factor
K(Fuel)[t]
fuel quality correction factor
K(Alt)[t] altitude correction factor

Fig. 3. GPS unit in a test vehicle.

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

Table 3
Average utilization and fuel efciency of different vehicles.
Mode: fuel

Vehicle utilization Fuel efciency No. of starts

(km)
(km/L)
(per day)

2-wheeler: petrol
21
Car: petrol
30
Car: diesel
30
3-wheeler: petrol
85
Taxi: gasoline
98
Taxi: diesel
98
Bus: diesel
186
Light commercial vehicles: diesel
45

48.8
10.1
9.5
21.5
10.5
10.1
3.5
8.6

9.1
7.8
7.8
14.4
10.8
10.8
11.7
6.7

K(Cntry)[t]
country correction factor
K[dt]
driving or soak style correction factor (also accounts for other
load effects from air conditioning usage and road grade).
As part of this study, Global Warming Potential (GWP) of different
pollutants was also calculated. GWP is one way to assess the potential
climate change impacts associated with emissions. GWP for each
pollutant was introduced by the IPCC (Intergovernmental Panel on
Climate Change) to determine the relative ability of GHG may have
towards forcing the earth's climate. GWP is based on the radioactive
efciency and decay rate of each gas relative to that of CO2 (Solomon
et al., 2007). The values of GWP vary as the concentration of GHG in
the atmosphere changes. GWP values for different pollutants for
100 year period are shown in Table 2.
4. Data collection
Driving pattern is an important factor in estimating vehicular
emissions. There are many methods to collect driving patterns including
the chase-car technique, private drivers drive their own vehicles with
instruments, private drivers drive instrumented vehicles, and private
vehicles with instruments are driven by professional drivers (Andre,
1996). We selected the latter method private vehicles with
instruments driven by professional drivers for this study. Randomly
hired local professional drivers drove instrumented vehicles. The drivers
were familiar with the roads and local trafc rules. The purpose of the
trip was not disclosed to drivers to avoid inuencing their behavior.
They were instructed to travel from point A to point B to collect secondby-second trafc conditions. Drivers were also asked to leave vehicles
on at trafc lights, since the GPS units were powered through the
vehicle's electrical system. Six vehicles of different makes, years, engine
sizes, and weights and three handheld GPS units were used. GPS units
were placed on the dashboards of vehicles to enable them to receive the

1803

best signal from the satellite as depicted in Fig. 3. Accuracy of the GPS
reading was ranging from 5 m to 25 m. Using the spatial analysis method
in Arc map software, spatial errors of GPS reading were corrected. This
has improved the accuracy of the GPS reading.
Instrumented vehicles were driven in trafc ow to record secondby-second positional information and trajectory. Drivers either accelerated or slowed down their vehicles depending on trafc conditions. The
test vehicle depicted the actual trafc prole of the link, since it traveled
at the representative trafc speed. In other words, it represented the
average speed of all types of vehicles such as two-wheelers, threewheelers, cars, and vans except buses on that link. This assumption is
considered valid for all modes including that for two-wheelers because
Chennai has heterogeneous trafc conditions, where about 20 types of
vehicles ranging from buses to bullock carts share the same road space
without any dedicated right-of-way. Buses were excluded since they
have different static and dynamic characteristics due to on-board and
drop of passengers at regular interval. Instruments were placed in buses
operating along different routes with the special permission of the
Metropolitan Transport Corporation (MTC) to collect the trajectory.
Data were collected in September 2004 during weekdays. There
were no unusual conditions such as major processions, VIP visits, or
other activities that could induce abnormal trafc characteristics in the
selected corridor during the survey. Instrumented vehicles were driven
during the morning peak (7:309:30 a.m.), afternoon off-peak (12:30
2:30 p.m.) and evening peak periods (57:30 p.m.). Technicians noted
street names, road surface characteristics, and causes for delays such as
pedestrian crossings, incidence of parking, and turning vehicles. Data
were generated for 276,000 s using cars and for 64,800 s using buses.
Data were downloaded and the collected data were quality-checked for
errors and inaccuracies. Driving patterns for different corridors were
developed with the MATLAB program. It has found that irrespective of
road classes, about 30% of time is spent below 20 kmph speed. Further, it
has revealed that current driving cycle (India Driving Cycle) doesn't
represent the real-world driving. For further details about the developed
driving pattern, please refer to Nesamani and Subramanian (2006).
4.1. Questionnaire survey
A questionnaire was developed and a primary survey was conducted
at ve fuel stations in the study area to elicit opinion on vehicle utilization
and starting patterns. Enumerators were trained on the conduct of the
survey and appraised of the purpose and signicance. The questionnaire
was tested before data collection. The survey was conducted from 6 a.m.
to 8 p.m. and about 495 usable samples were collected.
Information on average distance traveled per day, odometer
readings, vehicle models, fuel types, fuel economy, number of starts

Fig. 4. Soak time distributions of vehicles in Chennai.

1804

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

emissions can be broadly classied as cold start and hot start. They
depend on the number of hours vehicles are turned off. Cold starts emit
higher CO and HC levels when the engine temperature is below 400 C.
Further, it consumes more fuel than a warm start (Al-Hasan, 2007).
None of the previous studies have quantied the start emissions in
India. Fig. 4 shows the start time distribution of vehicles in Chennai.
5. Emission estimation analysis

Fig. 5. Estimated VKT from different vehicle classes.

per day, and off-time between starts were collected. Table 3 shows the
vehicle utilization, fuel efciency and number of starts for different
vehicles. Start patterns are critical for emission inventory. Start

Based on this methodology, emissions were estimated for local

and global pollutants using the IVE model for 2005. The estimated
emissions from motor vehicles in Chennai in 2005 were 431, 119, 46,
7, 4575, 29, and 0.41 tons/days respectively for CO, VOC, NOx, PM, CO2,
CH4 and N2O. About 19% of emissions are from start emissions. Figs. 5
and 6 illustrate the estimated travel demand and source of emissions
from different vehicle classes respectively. Local pollutants are of

Fig. 6. Estimated emission from different vehicle classes.

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

immediate concern for urban air quality since they have signicant
impact on health. Temporal distributions of emissions are required to
quantify the health impact of mobile source emissions in a city.
Previous studies have quantied the temporal distribution based on
trafc intensity and assumed the same for all pollutants. However, it
varies based on many factors such as road geometry, trafc characteristics, and weather characteristics. Fig. 7 shows the estimated daily

1805

temporal variation for different pollutants using IVE model. It can be

seen that start emissions are high during the morning peak hours.
Global warming potentials of direct and indirect effects of different
GHG pollutants were calculated using the values in Table 2. It is
estimated that on-road transport contributes about 6637 tons/day CO2
equivalent in Chennai. Fig. 8 illustrates the contribution of different
pollutants towards the GWP.

Fig. 7. Temporal variations of mobile source emissions in Chennai City.

1806

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

Fig. 8. Global warming potential of different GHG pollutants in Chennai.

In Chennai, about 64% of the total CO emissions are contributed by

two and three-wheelers, which primarily use two-stroke engines.
Stringent emission norms have reduced emissions per vehicle over the
last decade but the number of vehicles has increased dramatically
during the same period. About 18.2% of total CO emissions are emitted
during the start-up. Gasoline vehicles alone contribute about 88% of VOC
emissions. This is mainly due to lower combustion efciency in gasoline
engines, more specically in two-stroke engines. It has been estimated
that 1525% of two-stroke engine exhaust is unburned fuel (Pundir,
2000). In Chennai, 26% of VOC emissions are due to start-ups.
Heavy-duty vehicles (buses and trucks) contribute more than 60%
and 36% of the NOx and PM emissions respectively. A recent study
observed that the ground-level ozone in Chennai has reached 53 ppb,
which is unhealthy for sensitive groups (Pulikesi et al., 2006).
Currently, about 11% of PM emissions are from start-up emissions.
Private vehicles are a large and growing source of CO2. Recent
statistics show that overall fuel consumption within the transport
sector in India is increasing. Regardless of emission control, CO2
emission increases as fuel consumption increases. In Chennai, more
than 50% of CH4 is emitted from cars and about 70% of N2O is from
heavy-duty vehicles.
6. Strategies to control automobile emissions
The analyses of emission inventory clearly establish that air quality
in Chennai has degraded. This was further supported by time series of
ambient air quality data. The following three strategies have been
identied to reduce local and global pollutants in Chennai. Each
strategy is discussed in detail with reference to Chennai.
1. Reduction of Vehicle Kilometer Traveled (VKT) through transitoriented development and Travel Demand Management (TDM).
2. Adoption of advanced technology (vehicles, fuel and I/M program)
3. Transport System Management (TSM).
6.1. Reduction of VKT through transit-oriented development
Chennai has relatively high population densities and traditionally have mixed-use. However, the average trip length has in-

creased from 4 km to 11 km mainly due to urban sprawl over the

last four decades. In 1970 Chennai city was about 80 km2. This
has increased to 174 km2 in 2006. The population density has an
inverse correlation with the distance from the Central Business
District. It is a matter of common observation that as one moves
out of the core city, a mixture of land uses and isolated developments, reecting characteristics of urban sprawls are widely prevalent. This is due to lack of integrated planning, settlements in
the urban fringe heavily depend on the mother city, and hence
transportation demand increases signicantly. Further, total daily
person-trips in Chennai are projected to increase from 10.6 to
20.8 million in the next two decades (CMDA, 2007). Therefore,
the Chennai Metropolitan Development Authority (CMDA) plans
to intensify development along transit corridors through increased
FSI (oor space index) and relaxed parking regulations. A study
by L&T Ramboll has recommended increasing the FSI from 1.5 to 2.0
on both sides of the corridor to improve the patronage of MRTS.
Higher urban density can signicantly reduce VKT and increases the
share of transit use. Currently the MRTS carries only about 76,000
commuters per day and this is only one ninth of the projected share
of patronage. This may change once the third and fourth phases of
the MRTS corridor are completed and integrated with other public
transits.

6.2. Reduction of VKT through Travel Demand Management (TDM)

The modal shift from private vehicles to public transportation can
signicantly reduce VKT. Chennai City has 3000 buses that cater to
4 million passenger trips per day. The city has four rail corridors with
a combined length of about 140 km and that carry about 0.7 million
trips per day (Subramanian, 2008). The modal split between public
and personal transports (share of modes) has been about 35:65.
However, Second Master Plan (SMP) has ambitiously proposed to
increase the share of public transport from 35 to 70 by 2026 with a
sub-modal split of 60:40 between bus and rail. This could be possible
only through radical and far-reaching decisions.
Trafc management in Chennai focuses on TSM such as one-way
streets, turning restrictions and trafc control devices completely
ignore TDM. The underlying principle of TDM is to restrain extravagant
use of low-occupancy private cars and two-wheelers. The techniques
could be parking control, tolls, road pricing, motorist trafc restraint,
and staggering ofce and school hours.
The Mass Rapid Transit System (MRTS), the rst of its kind in the
country, was extended to about 20 km during 2007 in Chennai. Fleet
strength of the MTC was substantially augmented, for the rst time in
30 years. A further expansion, through JNNURM (Jawaharlal Nehru
National Urban Renewal Mission) funding, is under consideration.
However, mere promotion of public transport (PT) is inadequate to
shift commuters to PT from personal transport. Simultaneous disincentives for private vehicle use are required to achieve increased
ridership. Therefore, adoption of TDM techniques is a crucial emissioncontrol strategy.

Table 4
Implementation date of Bharat Stage norms and permissible sulphur level in fuel specications for different category of vehicles in Chennai.
Source: GOI, 2002; SIAM, 2009.
Modes

Fuel specication maximum sulphur % (w/w)

Effective date of mass emission standards

BS-I

BS-II

BS- III

BS-IV

BS-I

BS-II

BS-III

4-wheelers

April/2000

April/2001

April/2005

April/2010

2 and 3 wheelers

April/2000

April/2005

April/2008 to
April/2010

0.1 (gasoline)
0.25 (diesel)
0.1 (gasoline)
0.25 (diesel)

0.05
(gasoline and diesel)
0.05
(gasoline and diesel)

0.015
0.035
0.015
0.035

BS Bharat Stage norms.

BS-IV
(gasoline)
(diesel)
(gasoline)
(diesel)

0.005
(gasoline and diesel)
0.005
(gasoline and diesel)

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

6.3. Advanced technology

Developing advanced technology to meet stringent standards
may be a better way to reduce vehicular emissions. Bharat Stages I,
II, and III were introduced in 2000, 2001, and 2005 respectively for
four-wheeled vehicles in Chennai. Bharat Stage IV is expected to be
implemented in 2010. Bharat Stages I and II were introduced during
2000 and 2005 respectively for two- and three-wheelers and are the
most stringent norms worldwide.
Fuel quality must align with advanced technology to meet stringent
emission standards. Sulphur content in fuel is the most important
factor and has a drastic impact on emission-control technologies.
Table 4 indicates the required sulphur content to meet the different
levels of stringent emission standards. Fuel injection technology with a
two-way oxidation catalyst must be used to meet the Bharat Stage III
standards in two- and three-wheeler segments (Iyer and Badami,
2007).
Battery-powered cars and hydrogen-fuel-cell buses are alternative
technologies to signicantly reduce local and global pollutants.
However, the cost of such technologies is beyond the affordability of
most road users in India. In developed countries like the U.S., these
technologies are at demonstration stages and mass production is not
expected until beyond 20252030 due to performance limitations and
safety requirements (Walsh et al., 2007). In Chennai, the Department
of Transport has introduced about 5000 liqueed petroleum gas
(LPG)-based three-wheelers to promote alternative-fuel vehicles
(CMDA, 2007). Such technologies could be adopted at ecologically
sensitive areas in a limited way.
It has been well-established that properly designed and operated
inspection and maintenance (I/M) programs can signicantly reduce
vehicle emissions. To evaluate the benets if I/M program, Society of
Indian Automobile Manufacturers (SIAM) have organized pollution
control camps in major cities across India. They found a 3% reduction in
CO, 40% reduction in HC and 1520% improvement in fuel economy
(SIAM, 2001). In Chennai, there is no full-edged I/M program.
However, there is a mandatory requirement for all vehicles to undergo
a Pollution Under Control (PUC) check every six months. There are
about 66 Emission Test Centers in the city. Most are located at fuel-lling
stations. Because of lax enforcement, only 10% of vehicles undergo the
mandatory PUC which is ineffective in identifying the major polluters
(Zubeda, 2007).
6.4. Transport system management
Sound trafc system management principles coupled with modern
technology play crucial roles in improving trafc ow and road safety

1807

without impairing environmental quality. However, trafc management in Chennai is marked by inefciency, driver non-participation,
and corruption. Use of modern techniques (3D Calculation of
Stopping-Sight Distance from GPS Data; Area Trafc Management
System viz., SCOOT/SCAT; Advanced Passenger Information System) is
still in its infancy. Trafc control devices (manual control signals) are
outdated. They are inappropriately placed, inconsistently operated,
and improperly maintained. Road markings do not retain their
legibility and visibility. Nesamani and Subramanian (2006) found
that average speed in Chennai city is less than 20 kmph during peak
hours.
Parking is the most challenging problem in Chennai. Commuters
do not normally pay parking fees. Shoppers either indiscriminately
park wherever they nd space and pay (or ignore) a nominal fee. The
city has about 5100 passenger car equivalency (PCE) compared to
13000 PCE during rush hour (CMDA, 2007).
7. Scenario design
Scenarios are developed to address the complexity and uncertainty in forecasting the outcomes. To develop these scenarios, the author
analyzed historical data and examined various options and strategies.
Three scenarios were developed to assess the emission-reduction
prospects of policies in Chennai. The impact of the identied strategies
was quantied for 20052025. Scenarios in this analysis are based
on different stages of development in reducing VKT, penetration of
advanced technology, and trafc management. This will guide policymakers to develop and implement better initiatives.
7.1. Business-as-usual (BAU) scenario
In BAU scenario, it is assumed that vehicular growth will continue
as expected. Further, this will capture the change in modal split,
technology penetration, and utilization of vehicles. All new vehicles
will meet Bharat Stage emission norms as based on current policy
measures. However, it assumes there are no considerable changes in
transport or environmental policies. This is the reference case for
comparison purposes.
7.2. Intermediate (IM) scenario
The travel demand for passengers and goods will follow the same
growth curve as in BAU. However, this scenario allows for greater
changes by considering earlier penetration of advanced technologies,
increasing the share of mass transit, and introducing trafc-management schemes to feasibly reduce VMT.

Table 5
Policy options to control automobile emissions under different scenarios.
Policy options

Business-as-usual (BAU)

Intermediate (IM)

Enhanced (EH)

Reducing vehicle
kilometer traveled

The share of public transport will

continue with existing trend.

50% of passenger travel demand

will be met by public transport
(2.4 million VKT by buses).

Advanced technology

Penetration of new vehicles

conforming to Bharat Stage norms.
5% of vehicle travel demand will
be met by alternative fuels.

Penetration of new vehicles

conforming to Bharat Stage norms.
10% of vehicle travel demand will
be met by alternative fuels.

Transport system
management

Introduction of hefty parking

charges for 2- and 3-wheelers in
commercial areas.

Introduction of parking charges for

work trip which will reduce 1% of
vehicle travel demand.

Doubling the density along the MRTS corridor

which will reduce 3% of travel demand.
Modal split between public and private transports
will be 70:30 (about 3 million VKT by buses).
Earlier start date for the penetration of advanced
technologies.
10% of vehicle travel demand will be met by
alternative fuels.
5% of vehicle travel demand will be met by zero
emission vehicles.
Introduction of effective inspection and maintenance
program for in-use vehicles.
Replace pre-2000 trucks with better after-treatment
technologies (particulate traps).
Introduction of congestion pricing and increasing
the parking charges which will reduce 3% of vehicle
travel demand.

1808

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

Table 6
Change in modal split between scenarios.
Mode

Private
IPT
Public

Business-as-usual

Intermediate

2005

2015

2025

2005

2015

2025

Enhanced
2005

2015

2025

54.6
9.8
35.6

60.7
5.2
34.1

66.7
4.2
29.1

54.6
9.8
35.6

53.0
4.2
42.8

46.7
3.3
50.0

54.6
9.8
35.6

44.0
3.2
52.8

28.0
2.0
70.0

Table 7
Penetration of technologies by mode under alternative scenarios.
Mode: technology: fuel

2-wheeler: 2 stroke petrol

2-wheeler: 4-stroke: petrol
Car: gasoline
Car: diesel
Car: LPG
Car: hybrids
Car: BOV
3-wheeler: 2-stroke: petrol
3-wheeler: 4-stroke: petrol
3-wheeler: LPG
3-wheeler: BOV
Taxi: gasoline
Taxi: diesel
Taxi: LPG
Taxi: BOV
Bus: diesel
Bus: FCV
Goods

2005

85
15
74
24
2
0
0
90
10
0
0
30
70
0
0
100
0
100

2025
BAU

Intermediate

Enhanced

70
30
60
23
17
0
0
50
20
27.5
0
25
50
25
0
100
0
100

40
60
35
15
35
15
0
10
30
60
0
10
20
70
0
97
3
100

0
100
0
0
50
40
10
0
0
87.5
12.5
0
0
80
20
92
8
100

7.3. Enhanced (EH) scenario

The third scenario includes feasible/expected policies and also
strategies that will be difcult to implement. The assumptions are
similar to the IM scenario, but the penetration of advanced technologies and modal shift toward public transport will be earlier than
in IM. This scenario considers maximum options to reduce local and
global emissions from vehicles, which can be implemented by
aggressive government initiatives and people's participation.
Scenarios portray the range of future possibilities and illustrate the
scale of options needed to restrain growth in vehicle use, local
pollutants, and GHG emissions. The IVE model was used to evaluate
these three scenarios in which it was assumed that driving patterns

and start time distributions would be the same as in 2005. Table 5

suggests policies to reduce emissions and each scenario is built on the
previous one, but assumes progressively faster changes.
Table 6 shows the change in modal split among different scenarios.
In the BAU scenario, the share of public transport decreases from 36%
in 2005 to 29% in 2025. It is projected that 70% of passenger travel
demand will be met by public transport in 2025 (including MRTS) in
an enhanced scenario. These critical assumptions made in the paper
for the modal split and strategies adopted to achieve it have been
adopted based on the Second Master Plan for the city. The Master Plan,
as the name itself implies is the rst plan of the city and all sector
plans prepared by different functional agencies have to be within the
framework of the Master Plan. These assumptions are apparently very
ambitious and may not be realistic based on the current trends.
However, the modal split could be achieved by increasing frequencies
of public transits, and integrating bus transport with rail transport.
Travel Demand Management measures such as congestion pricing,
and increasing parking charges may also help in the augmentation of
the share of public transport. The emission estimation analysis is for
on-road transport only and does not include the rail system. The
expected increase in rail system usage will obviously reduce the total
on-road travel demand.
New technologies within each mode must be the state of the
solutions to meet stringent emission norms. This is based on discussions
with industry experts and on government plans and programs for
introducing new technology-based vehicles to improve air quality.
Table 7 gives the penetration of improved technologies under the
private, IPT, and public transport modes during 20052025 in the city.
Average speed is an essential input for emission estimation.
However, there is no reliable method available to predict the future
speeds. Further, there is no evidence based on the existing literature
to establish that trafc speed will go up or down in Chennai. Therefore, the average speed for the future is assumed to be the same as in
the base year. It is also assumed that speed limit for different vehicle
classes will not change in the future.

8. Policy evaluation
8.1. Travel demand
Travel demand for passenger and freight vehicles was estimated
for 20052025 with 10-year intervals using Eq. (1). The share of
vehicle travel demand provided by different modes is shown in Fig. 9.
The total VKT in Chennai is expected to double before 2015 and

Fig. 9. Estimated vehicle travel demand by different modes.

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

1809

Fig. 10. Share of travel demand by different modes in 2025 in different scenarios.

increase 5 times by 2025. The annual growth rate of private vehicles is

predicted to be about 21% between 2005 and 2025. This rapid growth
is attributable to an increase in per capita income, increased access to
passenger vehicles, and greater distances traveled. The number of
buses is less than 1% of vehicles, which caters to about 35% of the
passenger travel demand. However, in the future, it is expected to
reduce to 29% by 2025. The total freight demand in the city went up
6.4% annually from 2005 to 2025.
The impact of different policies on vehicle travel demand in 2025
in different scenarios is illustrated in Fig. 10. The share of private
vehicles has declined by about 32% and 60% in intermediate
and enhanced scenarios compared to the BAU scenario. On the
other hand, the share of public transit increased by about 42% and
71% /INS; in an intermediate and enhanced scenario respectively to
meet the growing demand in the city. The major factors that signicantly reduced the travel demand in an enhanced scenario were
due to densication of MRTS corridors and augmenting the public
transport.

8.2. Emissions
Fig. 11 illustrates the estimated emissions for three scenarios using
the proposed method. The rate of increase has reduced for all the
pollutants; however, the total emission load increased compared to
the 2005 level. In the BAU scenario, local pollutants dropped from
601 tons/day to 583 tons/day in 2015 and increased again to 758 tons/
day in 2025. In the case of global pollutants, there was a steep increase
in emissions from 4575 tons/day to 14,954 tons/day in 2025, which
might be due to rampant growth. This indicates that the proposed
policies have only a short-term effect on emissions and growth in
demand supersedes the benets. On the other hand, intermediate and
enhanced scenarios show a reduction in pollutants of 19% and 38% in
the next two decades respectively. This is due to strategies such as
penetration of advanced technology and trafc-management
schemes. This indicates that there is a potential to reduce emissions
in Chennai.
The impact of different policies on each pollutant has been
analyzed in detail. The intermediate and enhanced scenarios reduced
CO emissions by about 21% and 37% respectively in comparison to the
BAU scenario. This is mainly due to increased use of public transit and
an effective I/M program. In the enhanced scenario, HC reduced about
40% over 20 years. These changes are because substantial shares of
passenger vehicles are expected to shift toward alternative fuels. NOx
emissions drop below the 2005 level in 2025 in the enhanced

scenario, mainly due to the penetration of new and efcient technology, and the replacement of older, heavy-duty trucks.
PM emissions would increase more than two times over the next
two decades if existing trends continue. The phasing out of diesel
vehicles and introduction of after-treatment technologies have
reduced 20% and 49% of PM emissions in the intermediate and
enhanced scenarios respectively. Under the intermediate scenario,
CO2 emissions increase and the rate of increase start falling gradually
later. Even in the enhanced scenario, CO2 emissions increase primarily
due to high growth rates in total VKT. This suggests that this might
be due to insufcient penetration of fuel-efcient and clean-fuel
vehicles.
9. Discussion and conclusions
This study has quantied both local and global pollutants using the
IVE model. The proposed methodology captures the vehicles' modal
activities to estimate vehicular emissions and to illustrate the
temporal distribution of start and running emissions in Chennai.
Further, it analyzes the implications of various policy options for
reducing vehicular emissions under three alternative scenarios in
Chennai City. This study concludes that, under the ongoing economic
growth in Chennai, aggressive efforts are required to reduce vehicular
emissions. Central, state, and local governments should adopt the
following measures:
(i)
(ii)
(iii)
(iv)

Enhance and extend public transport infrastructures

Accelerate the adoption and deployment of improved technology
Deploy effective I/M programs
Implement optimized programs for trafc management and
enforcement.

Existing services and amenities should be improved by increasing

service frequency, reducing headways, improving service reliability,
adding bus shelters, properly designing trafc control devices, and
improving safety and cleanliness. Existing fare and pricing strategies
should be changed to support innovative methods such as transfer
policies and partnerships with businesses or other institutions to
provide discount fares. Train and bus services compete with, rather
than complement, each other. They should be integrated with in the
ambit of public transit agency as a rst step and later integrate with
private modes.
India should reduce sales taxes and import duties through
incentives to encourage faster penetration of advanced technologies
such as hybrid vehicles, BOV, and fuel cells. It is necessary to improve
the fuel quality to take advantage of after-treatment technologies

1810

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

Fig. 11. Estimated emissions under different scenarios.

such as particulate traps and catalytic converters. Since the Indian

transportation sector consumes more than 90% of HSD, it is imperative
to reduce the sulphur and benzene content in diesel fuel all over India.
Recently it has been permitted to use ethanol blend up to 5% in
gasoline (GOI, 2002).

Extensive application of technology and minimizing human interactions in enforcing regulations are reliable ways to improve the
situation. A vital aspect of enforcement is the manner in which individual police behave while enforcing trafc rules. Violators should
not be treated on par with criminals and offenders. Perhaps trafc-

K.S. Nesamani / Science of the Total Environment 408 (2010) 18001811

management responsibilities should be shifted from the police and

entrusted to a dedicated trafc-management team.
Acknowledgements
The author would like to thank the Ford Foundation International
Fellowship program (IFP) that provided the support to conduct this
research. Author would also like to thank Prof. K P Subramanian and
anonymous reviewers for their valuable comments on the earlier
version of this paper. The views expressed in this paper are of the
authors and are not necessarily the views of the IFP or the University
of California, Irvine.
References
ADB. Energy Efciency and Climate Change Considerations for on-road Transport in
Asia; Philippines: Asian Development Bank; 2006: 68. Available at < www.
cleanairnet.org/caiasia/1412/articles-70656_draft.pdf>, visited on February 8,
2009.
Al-Hasan M. Evaluation of fuel consumption and exhaust emissions during engine
warm-up. Am J Appl Sci 2007;4(3):10611.
Andre M. Driving cycles developments: characterization of the methods. Soc Automot
Eng 1996:961112.
Anh K, Rakha H, Trani A, Van Aerde M. Estimating vehicle fuel consumption and emissions
based on instantaneous speed and acceleration. J Transp Eng 2002;128(2):18290.
Badami MG. Transport and urban air pollution in India. Environ Manag 2005;36(2):
195204.
Bose RK, Nesamani KS. Issues, strategies, and challenges for sustainable urban
transport: modelling the impact on the local and global environment case of
Delhi. Pac Asian J Energy 2001;11(1):3149.
Chapman L. Transport and climate change: a review. J Transp Geogr 2007;15(5):
35467.
CMDA. Second Master Plan for Chennai Metropolitan Area. Chennai Metropolitan
Development Authority, 2007. Available at < http://www.cmdachennai.gov.in/
SMP_main.html > visited on January 15, 2009.
CRRI (Central Road Research Institute). Urban Road Trafc and Air Pollution (URTRAP).
Final report submitted to Centre for High Technology, Ministry of Petroleum and
Natural Gas, New Delhi, India; 2002.
Davis N, Lents J, Osses M, Nikkila N, Barth M. Development and application of an
international vehicle emissions model. Transportation Research Board 84th Annual
Meeting, Washington D.C.; 2005.
Energy Information Administration (EIA). International Energy Annual 2006. Available
at < http://www.eia.doe.gov/iea/carbon.html > visited on January 22, 2009.
Government of India (GOI). Report of the Expert Committee on Auto Fuel Policy.
New Delhi; 2002.
Gurjar BR, Van Aardenne JA, et al. Emission estimates and trends (19902000) for
megacity Delhi and implications. Atmos Environ 2004;38:566381.
IVE (International Vehicle Emission). IVE model user's manual version 2.0. 2008.
Iyer NV, Badami MG. Two-wheeled motor vehicle technology in India: evolution
prospects and issues. Energy Policy 2007;35:431931.

1811

Joumard R, Jost P, Hickman J, Hassel D. Hot passenger car emissions modelling as a

function of instantaneous speed and acceleration. Sci Total Environ 1995;169(13):
16774.
Karkezi S, Majoro L, Johnson TM. Climate change mitigation in the urban transport
sector: priorities for the World Bank. Washington D.C.: The World Bank; 2003
Kokaz K, Harshadeep NR, Rogers P, Srinivasan S. A simulation analysis of transportation
policies on health and environment in Delhi, India. Computers in Urban Planning
and Management Conference, University of Hawaii; 2001.
Mittal LM, Sharma C. Anthropogenic emissions from energy activities in India:
generation and source characterization. Emissions from vehicular transport in
India (part II). Ohio, Ohio Supercomputer Center; 2000.
Nesamani KS, Subramanian KP. Impact of real-world driving characteristics on
vehicular emissions. JSME Int J Ser B 2006;49(1):1926.
Pachauri RK, Sridharan PV, editors. Looking back to think ahead, GREEN India 2047.
New Delhi: Tata Energy Research Institute; 1998.
PTCS (Pallavan Transport Consultancy Services). The comprehensive Trafc and Transportation Study (CTTS), Chennai, Chennai Metropolitan Development Authority;
1995.
Pucher J, Peng Z, Mittal N, Zhu Y, Korattyswaroopam N. Urban transport trends and policies
in China and India: impacts of rapid economic growth. Transp Rev 2007;27(4):
379410.
Pulikesi M, Baskaralingam P, Elango D, Rayudu VN, Ramamurthi V, Sivanesan S. Air
quality monitoring in Chennai, India, in the summer of 2005. J Hazard Mater B
2006;136:58996.
Pundir BP. Vehicular air pollution in India: recent control measures and related issues.
In: Morris S, editor. India Infrastructure Report 2001. New Delhi: Oxford University
Press; 2000. p. 2603.
SIAM. SIAM viewpoints for improved inspection & maintenance in India. GITE Regional
Workshop on Transport Sector Inspection & Maintenance Policy. 10 12 December;
2001. Bangkok.
SIAM, Emission Norms, Available at < http://www.siamindia.com/scripts/emissionstandards.aspx > visited on January 10, 2009.
Sengupta B. Road map for fuel quality improvements to reduce air pollution in India.
PETROTECH 2005, January 1718; 2005. New Delhi, India.
Sjodin A, Lenner M. On-road measurements of single vehicle pollutant emissions, speed
and acceleration for large eets of vehicles in different trafc environments. Sci
Total Environ 1995;169(13):15765.
Solomon S, et al. Climate change 2007: the physical science basis, IPCC Fourth
Assessment Report (AR4). Cambridge, United Kingdom and New York, NY, USA:
Cambridge University Press; 2007. 996 pp.
Srinivasan S, Rogers P. Travel behavior of low-income residents: studying two
contrasting locations in the city of Chennai, India. J Transp Geogr 2005;13:26574.
Subramanian KP. Public transportation system in Chennai, India: a review. Municipalica
making cities work better, Mumbai, India; 2008.
The World Bank. For a breath of fresh air ten years of progress and challenges in
urban air quality management in India (1993 2002). Environment and Social
Development Unit, South Asia Region. India: The World Bank; 2005.
Walsh MP, Kalhammer FR, Kopf BM, Swan DH, Roan VP. Status and prospects for zero
emissions vehicle technology. Prepared for State of California Air Resources Board,
Sacramento, California; 2007.
WSA (Wilbur Smith Associate). Study on trafc and transportation policies and
strategies in urban areas in India. Ministry of Urban Development India; 2008.
Zubeda H. Emission testing falls, so does number of centres, The Hindu, Tuesday, Nov 13,
2007. Available at< http://www.hindu.com/2007/11/13/stories/2007111364450300.
htm > visited on February 10, 2009.