Assessment of interstate freight vehicle characteristics and impact of future emission and fuel economy standards on their emissions in India

Energy Policy 108 (2017) 121–133

Contents lists available at ScienceDirect

Energy Policy journal homepage: www.elsevier.com/locate/enpol

Assessment of interstate freight vehicle characteristics and impact of future emission and fuel economy standards on their emissions in India

MARK

⁎

Leeza Malik , Geetam Tiwari Transportation Research and Injury Prevention Programme, Indian Institute of Technology, Hauz Khas, New Delhi, Delhi 110016, India

A R T I C L E I N F O

A BS T RAC T

Keywords: Freight Emission standards Fuel economy India Highway traffic

Road freight in India plays a vital role in the economic growth of our country. Rising energy consumption and the environmental impacts by freight vehicles is constantly monitored through the introduction of stringent emission standards and through the promotion of fuel efficiency policies. However, little information is available on freight activity data (vehicle-kilometre) and their fleet characteristics (vehicle age, annual mileage, fuel economy etc.) to assess these policies. Therefore, the paper analyses the existing fleet characteristics in terms of age, annual mileage, fuel economy, fuel type used etc. for the freight vehicles used for “interstate” or “intercity” mobility on National Highways. Origin-Destination surveys and traffic volume counts are conducted at ten locations along major National Highways to capture the fleet characteristics. Based on the results, emission outlook for the proposed emission standards and fuel economy standards for freight vehicles is presented. Significant emission reductions are expected if emission standards are nationwide and timely implemented. Further, the requirement of country representative survival rates is highlighted.

1. Introduction The transport sector is globally recognised as the backbone of economic growth. However, gradually concerns have been raised for its close association with the energy consumption and its related environmental impacts. Globally, transport sector accounts for 23% of CO2 emissions. Within transport sector, 75% of CO2 emissions are shared by automobiles and trucks travel (IEA, 2009). Sustainable and strategic planning for the transport sector has, therefore, been a growing thrust worldwide. Being a developing economy, in India, similar challenges have been realised and various policies have been implemented to combat the rising energy consumption and environmental emissions by the transport sector. India is facing an unfortunate recognition with 33 Indian cities among top 100 world's worst pollution affected cities (WHO, 2016). In fact, the health cost from air pollution constituted 23% of the total India's GDP value in 2010 (TERI, 2015). Emissions and fuel consumption from the road transportation has always been a focus of concern and is rather constantly monitored through the introduction of emission standards and through the promotion of fuel efficiency policies. Nevertheless, these worthy attempts have been marred by the increase in population, rapid urbanisation, rise in motorisation and limited investments in sustainable transport systems (Pucher et al., 2005). According to Census of India (2011), the population of India increased to 1210 million registering a decadal

⁎

Corresponding author. E-mail addresses: [email protected] (L. Malik), [email protected] (G. Tiwari).

http://dx.doi.org/10.1016/j.enpol.2017.05.053 Received 20 February 2017; Received in revised form 13 May 2017; Accepted 26 May 2017 0301-4215/ © 2017 Elsevier Ltd. All rights reserved.

growth of 17.64%. Meanwhile, the number of registered motor vehicles increased to 210 million with a compound annual growth rate of 9.8% between 2005 and 2015 (MoRTH, 2016). Aggressive growth rate both in terms of population and vehicles have, therefore, continuously challenged the balance between India's energy demand, air quality and overall development. The transport sector in India contributes to 6.4% share in India's GDP of which road transport accounts for 4.5–5% share. In 1950s road transport carried 15% of passenger and 14% of freight movements with the total network length of 0.4million km. However, over the past two decades, road sector has evolved as a predominant mode of transport both in terms of the number of vehicles and road network length. This is evident from the fact that road transport now accounts for 86% of the passenger movements and 66% of freight movements with the total network length of 4.7 million km in 2011 (MoRTH, 2011) The rise in the share of road transport is attributed to its flexibility and adaptability in operation and is supported by massive investment in road infrastructure through government programmes like the National Highways Development Project (NHDP), Pradhan Mantri Gram Sadak Yojana (PMGSY) and Jawaharlal Nehru National Urban Renewal Mission (JnNURM). As far as energy consumptions are concerned, the transport sector of India accounts for 14% of the final energy consumption and has been associated with the highest growth rate in terms of energy consumption

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

impacts of accelerated phasing of fuel economy standards are also explored. The paper includes five sections. Section 2 discusses the timeline of fuel emission standards and fuel economy standards for freight vehicles in India. Section 3 deals with data collection methodology. Section 4 deals with the discussion of results and presents the emission outlook through years 2016–2026. Section 5 deals with policy implications and presents the way forward for the future work ahead.

(annually 6.8% since 2000) among other end user sectors (IEA, 2015). A significant component of the transport sector energy consumption is dominated by the oil-fuelled road travel by freight vehicles. PCRA (2013) also highlights that “Of the total diesel consumed by road transport, trucks and buses accounted for about 77% with buses consuming around 7.08 million tonnes per annum and trucks consuming 24.25 million tonnes per annum.” Further, Dhar and Shukla (2015) indicate a strong correlation between per capita freight demand and energy demand. The authors project the increase in freight demand with 8% GDP growth to increase from 1464btkm in 2010 to 5941btkm in 2050. Apart from the large proportion of oil consumption, freight transport has also attracted much concern due to the release of toxic air pollutants and greenhouse gases (GHGs). Ramachandra (2009) estimates that truck and lorries contribute to the highest proportion of vehicular emissions in India. Total emissions include CO2 (28.8%), NOx (39%), SO2 (27.3%) and PM (25%). On similar line, the emission inventory for different cities in India suggests that diesel operated heavy and light-duty vehicles have the highest contribution to the overall emissions (Goel and Guttikunda, 2015; Guttikunda and Kopakka, 2014; Guttikunda and Mohan, 2014). In spite of the vital role of road freight sector in the overall energy consumption and air emissions, the policy environment of freight vehicles in India has been quite sluggish in comparison to its counterpart—passenger transport. Road freight sector has seldom being involved at national level planning, eventually leading to some adhoc measures at the urban level to mitigate its negative impacts. For instance, the capital city of India, Delhi has introduced time and again different policies like time restriction, green tax and prohibition of freight vehicles greater than 15 years to curb city's air pollution. Besides these policy interventions, the benefit of these localised efforts still remains questionable. Therefore, at present, there lies a significant scope of nationwide efforts through the improvement of freight vehicle emission standards and fuel economy standards. Strict and timely implementation of these standards will consequently help to reduce fuel consumption, encourage technology advancements, reduce emissions and cut down energy demand. However, the paucity of freight activity data (vehicle-kilometre) and their fleet characteristics (vehicle age, annual mileage, fuel economy etc.) limits the analysis of the potential impacts of the key policy initiatives (CAI-Asia, 2011; Planning Commission, 2013). Unlike other developed countries, in India national travel or commodity flow surveys are not conducted on regular basis. The information is available through few occasional travel surveys conducted as a part of independent studies (Schipper et al., 2009). One of the attempts to estimate commodity specific interregional freight traffic was made in 2009 (RITES, 2009) where the main focus of the report was on estimating nationwide tonne-kilometre and its projections over years. No specific information regarding the freight characteristics is available through the study. Taking into account the above facts, the present paper aims to analyse the existing fleet characteristics in terms of age, annual mileage, fuel economy, fuel type used etc. for the freight vehicles used for “interstate” or “intercity” mobility on National Highways. This includes the freight vehicles used for long distance transportation. Secondly, comparisons are made between observed freight characteristics to that of urban fleet characteristics available for the city of Delhi. Various studies available at the city level or national level are referred to for comparisons (Baidya and Borken-Kleefeld, 2009; Goel and Guttikunda, 2015; Gurjar et al., 2004; IISD, 2013; Pandey and Venkataraman, 2014; Ramachandra, 2009). Thirdly, implications of various assumptions used in the literature regarding these characteristics on the total amount of emissions are investigated for the base year, 2016. Also, based on the observed characteristics three emission scenarios are tested—(a) Business as usual- No improvements in the present emission standards (b) BS IV emission standards are introduced nationwide in 2017 (c) BS VI emission standards are introduced nationwide in 2020 after implementation of BS IV in 2017. Further, the

2. Background fuel emission and economy standards - India 2.1. Fuel emission standards Emission standards are formulated to lower emissions by effectively designing new vehicles as compared to the old fleet. India uses the European Union based “technology following” emission standards. Through this approach emission levels are reduced by the use of already demonstrated technology (Faiz et al., 1996). In 1992, for the first time emissions from diesel vehicles came under jurisdiction. Next revision in the emission standards came in the year 1996 which was followed by the adoption of Bharat Stage I standards (Equivalent to Euro I) in the year 2000. Subsequently, there have been progressive efforts to reduce the emission levels through the adoption of stringent emission standards. In 2003, an expert committee was constituted to chart out the roadmap for the emission standards in India on regular basis. In 2014, the committee submitted its second report on “Auto Fuel Vision and Policy, 2025 (2014)″ and recommended to introduce BSV and BSVI emission standards by 2020 and 2025, respectively. However, realising the need for stricter emission standards the road ministry in 2016 decided to leapfrog directly from BSIV to BSVI emission standards nationwide by 2020. The emission standards for the freight vehicles is divided into two categories—light-duty vehicles (with gross weight less than 3.5 t; include all light commercial vehicles like three-wheeler, minivans used for commercial purpose and exclude passenger cars or minivans used to carry passengers) and heavy-duty vehicles (with gross weight greater than 3.5 t; includes medium and heavy duty trucks with 4-axle or more). Fig. 1 shows the reduction in the emission level of various pollutants (g/kWh) for diesel operated heavy-duty vehicles through the adoption of Bharat stage standards with time. The emission standards for heavy-duty vehicles are quite different from light-duty vehicles. Where the light-duty vehicles follow the standards set for the passenger vehicles, establishing the emission standards for heavy-duty vehicles is a complicated task. Emission standards for heavy-duty vehicles are reported in terms of grams of pollutant per unit power generated by the engine (Engine-based certification) instead of grams per kilometre (Vehicle-based certification). Vehicle-based certifications may be more accurate than engine based certifications due to the fact that engine performance is strongly affected by the type of chassis it is paired with, vehicle size and fuel consumption. On the other hand, vehicle-based standards would

Fig. 1. Chronology of emission standards for heavy-duty vehicles (g/kWh) in India.

122

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

the adopted structure various market distortions thus appeared—(a) Shift of station wagons to minivans and SUVs with lower fuel economy; (b) Emergence of large SUVs to comply with lower fuel economy targets set for light-duty trucks; (c) Advent of larger pickup trucks to escape from the CAFE weight restrictions. However, U.S. revised the CAFÉ standards and introduced vehicle footprint (product of the length of wheelbase and track width) based fuel economy standards in 2011. In simple words, the fuel economy targets that each manufacturer has to comply with were related to the footprint of vehicles through a mathematical equation. These attribute-based fuel economy standards, as opposed to uniform targets, intends to set differential targets for manufacturers depending on the size of vehicle mix sold, thus, introducing the degree of difficulty faced by small manufacturers to comply with targeted fuel economy standards (Plotkin, 2009). Fuel economy standards in the European Union started with the voluntary agreement with the automobile manufacturers in the year 1998. The fuel economy standards used CO2 reduction as the measuring unit and aimed at reducing average emission to 140gCO2/km by 2008 in comparison to the average of 187 g/km in 1995 for passenger cars (Zachariadis, 2006). Initially, light commercial vehicles (less than 2.61 t; commonly include vans) were not included under the bars of fuel economy standards. The adoption of CO2 reduction standards by the European Union instead of fuel consumption based standards was in view to its commitment to reduce greenhouse gas emissions under Kyoto Protocol (An et al., 2011). Falling short of the targets in 2008 the European Union adopted mandatory weight-based fuel economy standards to achieve 130 gCO2/km for passenger cars by 2015 (Schipper, 2011). In addition to this, light commercial vehicles CO2 emission targets were also adopted in 2011–12 that aimed at achieving the target of 175 g/km by 2016 and of 147 g/km by 2020 (DieselNet, 2017). Policymakers have to distinguish between fuel based and CO2 based standards, as both cannot be used interchangeably if different fuel types are used by the fleet. This is because there are differences in the carbon contents of different fuel type—not necessarily translating into similar CO2 impacts. Conversion factors are applied to convert the gasoline and diesel fuel consumptions to reflect CO2 equivalents. These conversion factors are assumed to vary marginally across various countries. However, the assumption of constant conversion factors may only hold true until and unless non-petroleum-based alternative fuels do not proliferate on widespread extent (Bansal and Bandivadekar, 2013). Japan adopted weight-based fuel economy standards in 1999 for passenger cars and light commercial vehicles (gross vehicle weight less than 2.5 t in both cases) and further revised them in 2007. The standards were based on “top runners” method. According to this method, efficiency standard for the target year is set above the most efficient model available in the base year (An and Sauer, 2004). The weight-based standards as compared to size-based standards (footprint based) are closely related to fuel economy, however, it limits the strategies to reduce the weight of vehicles while increasing fuel economy and encourages the use of heavier engines complying to lower fuel economy (Plotkin, 2009). To resist this tendency Japan government has set up progressive taxes on the purchase of heavier weight vehicles. Thus, from the discussion, it can be observed that adoption of fuel economy standards for light commercial vehicles dates long back and is accompanied by passenger cars fuel economy standards. On the other hand, heavy-duty fuel economy standards are a new endeavor worldwide with Japan being the first one to adopt them in 2005. As of now, only four countries, i.e., U.S, Japan, China and Canada have adopted the heavy-duty fuel economy standards. Other countries like India, Mexico and South Korea are also looking forward to introduce these standards sometime in the future. Further, the European Union is also monitoring and planning regulations for CO2 emissions from heavyduty vehicles (ICCT, 2016).

require several vehicle categories to be tested on chassis dynamometer individually, and thus several sets of emission standards. To avoid such intricacies and to have a broader range of applications engine based standards are therefore generally followed (Faiz et al., 1996). Engines are tested over different testing cycles to simulate on-road engine performance and these tests may vary from one country to another. As of now in India, European Stationary Cycle (ESC) and the European Transient Cycle (ETC) are used to test BS III and BSIV compliant vehicles. ESC is used to test engine over steady state cycles whereas ETC is used to test engine over different cycles of driving conditions. However, with the proposed introduction of BSVI standards in 2020 better estimates for emissions are anticipated through the use of World Harmonised Transient Cycle (WHTC) and the World Harmonised Stationary Cycle (WHSC) tests (Bandivadekar, 2015). 2.2. Fuel economy standards Fuel economy standards are adopted to check increasing energy dependency and, consequently, limiting the greenhouse emissions. Indian government finalised the first fuel economy standards for cars in the year 2014 and these standards came into effect on 1st April 2017. The draft of first fuel economy standards for freight vehicles is expected to be released sooner or later in the year 2017. Regulation of fuel economy standards for freight vehicles in India is strongly advocated given the largest potential reduction in fuel consumption among the major markets of heavy-duty vehicles around the world (ICCT, 2016). Policymakers encounter several questions while drafting new fuel economy standards of which three aspects play a critical role in setting the entire framework (Plotkin, 2009):

• • •

What are the optimum target levels Timing and structure of fuel economy standards Stringency and implementation method

Unfortunately, there is not one solution to these questions as different approaches have been adopted around the world depending on their historical experiences, cultural and political environment. Indeed, it makes sense to understand the effects of these different approaches by reviewing the past experiences and future considerations of prevalent fuel economy standards. Worldwide, fuel economy standards vary between countries based on following measures (Atabani et al., 2011) —(a) Fuel efficiency approaches (Fuel consumption reduction, CO2 reduction, GHG reduction); (b) Measure (mpg, g/ km, km/l, l/100 km, g/mile); (c) Structure (cars and light trucks, weight-based, overall light-duty fleet, engine size) (d) testing methods (U.S. CAFÉ, EU NEDC, Japan 10–15); (e) Implementation (mandatory, voluntary). Each of these measures has implications on the timely achievement of targeted fuel economy standards. To understand the pro and cons of the measures we review the fuel economy standards of three largest automobile markets of the world i.e., U.S., European Union and Japan. The first fuel economy standards were formulated by U.S. for passenger cars and light-duty trucks (include light commercial vehicles; vans; pickups; sports utility vehicles (SUVs); gross weight less than 3.855 t) in the 1970s due to the oil crisis (Atabani et al., 2011; Cheah and Heywood, 2011; Karplus et al., 2013; Plotkin, 2009). The fuel economy targets measured in miles per gram for the light-duty vehicle were initiated in 1979 and further revised in 2004. According to their Corporate Average Fuel Economy program (CAFÉ), manufactures were to mandatorily comply with sales-weighted fuel economy standards for all vehicles sold in each manufacturer's fleet. Also, it is to be noted that initially, the fuel economy targets for light-duty vehicles (20.2 mpg in 1991) were lenient in comparison to that for cars (27.5 mpg in 1991) (Atabani et al., 2011). Further, improvements in fuel economy targets were stagnated after 1995 particularly for lightduty vehicles (Schipper, 2011; Zachariadis, 2006). In consequence to 123

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

the total distance travelled in the survey week. This type of surveys has an advantage that a large sample size can be collected—however, the chances of underreporting by vehicle drivers cannot be ignored. The latter approach of calculating fuel economy involves limited sample lab tests that simulate on- road driving conditions. The limitations of this method lie in the fact that fuel economy estimates are to be extrapolated for different types and vintages of the vehicle. Secondly, it may not be able to simulate the variation in speeds experienced in a typical journey. Thirdly, for freight vehicles, these tests are performed by assuming an average load factor which may not hold true in real-world situations (McKinnon and Piecyk, 2009). It is clear from the above discussion that different data collection methodologies have different data requirements and different degree of errors associated with it. Indeed, emission scenarios will inevitably be affected by the data collection approach one chooses. However, by being aware of the underlying assumptions and by avoiding the potential errors in each data collection process, the accuracy of the baseline scenarios can be improved.

3. Methodology 3.1. Baseline data requirements The prediction of the potential impacts of the emission related studies relies on the soundness of the “reference scenario” against which other scenarios are compared with. The reference scenario reflects the present trends of various parameters affecting the given policy environment. Accurately establishing and understanding these trends may help to avoid misleading results and will, therefore, yield cost effective solutions (Zachariadis, 2006). This section, thus, will discuss various data collection approaches practised to capture information necessary for the analysis of emission standards and fuel economy scenarios. The information to be collected before implementation of any freight emissions reduction policies include exhaustive dataset regarding: Freight vehicle activity; fleet age; fleet turnover; fuel economy; freight market; payload information; vehicle maintenance practices; vehicle speed; and technological options available for emission reduction (Clark et al., 2002b; Hill et al., 2011; Huai et al., 2006; Zamboni et al., 2015, 2013). Freight vehicle activity refers to the annual kilometre driven by the vehicles on a road network. To accurately measure the annual kilometre travelled one may ideally require to gather travel data from each individual vehicle. However, conducting these types of surveys is in reality practically impossible and economically infeasible. Thus, two broad methods: a) estimates based on traffic counts, b) estimates that are not based on traffic counts, are used to calculate the vehicle activity (Kumapley and Fricker, 1996). The first approach is based on collecting annual average daily traffic (AADT) data on randomly stratified selected sample sections. Then, the vehicle travel on these sample sections is obtained by multiplying the observed AADT with corresponding road length. Further, these annual kilometre estimates are expanded to represent the area wide estimates (DOT, 2014). The primary shortcoming of this approach is the prior knowledge of traffic data on state network for stratification purpose. Perhaps, this may not be available in most situations. Although, stratification for sampling may be based on the other functional classes like the number of lanes, the accuracy of achieving stratified homogeneous groups may be low (Kumapley and Fricker, 1996). The second, non-traffic count based approach for determining the annual kilometre driven includes: a) Estimates based on fuel sales data b) Estimates based on Odometer recordings. The fuel sales based vehicle kilometre estimates require the information regarding the revenue generated by fuel sales, the unit price of fuel and the information regarding fuel economies of the vehicles (Williams et al., 2016). The degree of error associated with this method is quite high and thus, it should be used with caution. First, a substantial degree of uncertainty is introduced due to the fuel economy estimates, as this, itself is a function of vehicle age, vehicle maintenance practices, driving conditions etc. Secondly, it is very difficult to distinguish whether all the fuel sold in an area is used to travel in that area only (Kumapley and Fricker, 1996). Thirdly, fuel sales data are usually not distinguished by vehicle type at the point of purchase (McKinnon and Piecyk, 2009). Another approach based on odometer recordings is to collect samples based on the driver survey. This involves recording the odometer values at the start and end of one year to indicate the annual mileage. However, this approach may be quite cost-intensive and may give unreliable results if readings from tampered /un-calibrated odometers are used for analysis (Williams et al., 2016; Kumapley and Fricker, 1996). Another set of studies is related to the data collection procedures to capture fuel economy data for emission estimates (Apelbaum, 2009; McKinnon and Piecyk, 2009; Schipper, 2011; Zachariadis, 2006). This is also broadly divided into two methods: a) Survey-based estimates b) Vehicle test cycle estimates. In the former approach, the drivers are required to provide details regarding the litres of fuel purchased and

3.2. Data collection methodology The survey-based methodology was adopted to capture the freight fleet characteristics in the month of May 2016. The data was collected along the major National highways NH-8, NH-3, NH-6 and NH-2 which connects major cities of India like Delhi, Ahmedabad, Mumbai and Kolkata. In absence of any preliminary information about AADT, ten sections were selected which served to contain major merging or diverging traffic points from National Highways. This criterion was initially chosen to capture the variability in the characteristics of freight fleet (particularly, age) originated from and destined to different states. The data collection procedures included 24-h classified volume counts, 4-h video recording and 48-h origin and destination (O & D) surveys. O & D surveys were conducted at freight rest places, “dhabas”, lying nearby traffic count locations. Two reasons can be associated with choosing the off-road sample methodology: First, the roadside survey would have required the help of local law enforcement authority to pull out the vehicles from the traffic stream for the interview. This may have involved unnecessary cost and delay in the procedures. Secondly, onroad intercept surveys conducted at entry locations by Malik et al. (2015) in Delhi for freight vehicles highlight the problem of reporting wrong odometer readings by drivers. In order to avoid this discrepancy, face to face interviews at rest places provided an opportunity to the surveyor to observe the odometer readings by himself. Traffic counts for freight vehicles were classified into following categories: (i) 2-axle trucks (ii) 3-axle trucks (iii) Multi-axle trucks ( > 4 axles) (iv) Goods auto (Commercial three wheelers) (v) Mini light commercial vehicles (can be compared with small vans with gross weight less than 3.5 t). The first, second and third categories were combined to form heavy-duty vehicle (gross weight greater than 3.5 t). Fourth and fifth categories were combined to form light- duty vehicles (gross weight less than 3.5 t). The validity of the manual counts was cross-checked with the video counts available for 4 h. It was observed that in the case of freight vehicles the count at most differed by ten vehicles. Given the fairness and accuracy in data reporting, we used the recorded manual counts for setting sample size targets at each location. Targets were set to achieve at least 5% sample of traffic count each for heavy-duty vehicles and light-duty vehicles. However, the aim was to maximise the number of samples to be collected at each surveyed site. Table 1 shows the number of samples collected at each location for heavy-duty vehicles and light-duty vehicles as a percentage of the observed daily volume of heavy-duty and light-duty vehicles, respectively. It is observed that at a few locations light-duty vehicle is underrepresented as compared to the targeted value. This may be attributed to the fact that not many of the light-duty vehicles use the rest places. Thus, a total of 393 and 4921 samples were collected for light-duty and heavy-duty vehicle, respectively. Similar methodology 124

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

methodology includes following components—(a) Activity: refers to the vehicle-kilometre or tonne-kilometre; measures the variation of distance travelled over years (b) Share: refers to the share of a particular mode of transport in the total activity performed; measures the preference of one mode over another (c) Intensity: refers to the fuel intensity (fuel/kilometre); measures the effect of speed, congestion and driving conditions (d) Factors-refers to the emission factors or the carbon content of different fuel type. ASIF approach permits to quantify in detail the impacts of various local or national policies on each of these parameters, and thus allows the estimation of potential consequences of dynamic transport system on overall emissions rates. However, the methodology presents in itself a challenging task with a wide range of on-road vehicle age mix, road types and operating conditions. The emissions for the criteria pollutant, i.e., PM (Particulate Matter), NOx, and CO is calculated using Eq. (1).

Table 1 Percentage of samples collected at each location for heavy-duty vehicles and light-duty vehicles. Survey Locations

Heavy-duty vehicle (%)

Light-duty vehicle (%)

Location Location Location Location Location Location Location Location Location Location

11 13 5 3 20 19 55 14 18 43

17 6 3 0 11 29 19 3 3 8

1 2 3 4 5 6 7 8 9 10

for data collection of freight characteristics can be found in studies like Strategic Freight transportation O & D study conducted with the cooperation of Washington Department of transportation (M. L. Clark et al., 2002a), Texas truck data collection guidebook (Prozzi et al., 2004) and NCHRP report on freight transportation surveys (National Cooperative Highway Research Program, 2011). Origin and destination surveys included the following information: (i) Vehicle registration number (ii) Vehicle type (iii) Vehicle ownership (iv) Model year (v) Odometer reading (vi) Type of fuel used (vii) Fuel economy (km/l) (viii) Commodity type (ix) Commodity weight (x) Origin state (xi) Destination state (xii)Trip frequency. The first, second and fifth questions were observed by the surveyor and responses for rest other questions were captured by interviewing freight drivers. The samples collected through the surveys captured the fleet characteristics of some of the states accounting for the highest annual registration of the freight vehicles, like Maharashtra (24%), Uttar Pradesh (16%), Haryana (13%), Rajasthan (15%), Gujarat (9.4%), etc. To check if the selected ten locations were representative of the age of the fleet travelling on National Highways, a hypothesis was formulated to test whether there was a significant difference in the mean ages at these ten locations. The hypothesis is based on the assumption that fleet age distribution can potentially explain other fleet characteristics like fuel economy, annual mileage etc. (Goel et al., 2015). As the number of samples collected at each location was different and variances observed were non-homogeneous BrownForsythe test (instead of ANOVA) was used to check the equality of mean age. It is observed and expected that the mean age of fleet at the ten locations was not same in the case of heavy-duty vehicles. However, in the case of light-duty vehicles, the mean age at these ten locations was not significantly different. Further, to check the pairwise statistical difference between multiple mean age at different location GamesHowell post-hoc test was performed. Games-Howell is the robust test that takes into account the unequal variances and unequal group sizes when comparing multiple means. In the case of heavy-duty vehicles, it was observed that mean age at locations one, two and eight did not differ significantly at 95% confidence level. Likewise, third, fourth, fifth, sixth, seventh, ninth and tenth locations did not differ statistically in their mean ages for heavy-duty vehicles. As indicated earlier, light-duty vehicles did not have a statistical difference between mean ages at all locations. The results of the post hoc test also supported the prior observation. Through the results, we assume that in the case of heavyduty vehicle and light-duty vehicles, a further collection of data may lead to the mean ages closer to the mean age of the groups identified and thus, we discontinued our data collection process.

Ep, v, a, f = Vehicleinusev, a xSf xVKTv, a xEHp, v, a, f

(1)

where E = Emissions; p = pollutant; v = vehicle type; a = age group; f = fuel type; VKT = vehicle kilometre; EF = Emission factors; Sf = share of fuel type. The emissions for CO2 are based on the fuel economy of the vehicle and is given by Eq. (2).

Ep, v, a, f = Vehicleinusev, a xSf xVKTv, a xFEv, a, f xPp, f

(2)

where E = Emissions; p = pollutant; v = vehicle type; a = age group; f = fuel type; VKT = vehicle kilometre; FE = Fuel economy; Sf = share of fuel type; P = pollutant content. The variation of annual mileage, fuel economy with age and fuel share is obtained from sample surveys. Calculation of in-use vehicles in literature is based on vehicle sales and survival rates (Huo et al., 2007). Modified Weibull distribution, log-logistic, Weibull, lognormal distributions are typically used for survival analysis (Baidya and BorkenKleefeld, 2009; Goel et al., 2015; Lee and Wang, 2003; Zachariadis et al., 2001). Survival analysis is used to capture the probability of vehicle registered in a particular year to be present on road for a given year. The total on-road fleet (Nk,y) for a given year y and freight type k is given by Eq. (3). y −1

Nk , y =

∑ Rk,x *Sk,y−x x

(3)

where Rk,x refers to domestic sales of freight type k for the year x. Sk,y−x refers to the survival rate of freight type k for age y–x. In India, the newly purchased vehicle has to be officially registered. However, separate registration renewal procedures are then followed for passenger cars and freight vehicles. The passenger cars have to register once for 15 years and no records are further updated to deduct scrapped/retired/resold vehicles. This leads to the overestimation of actual on-road vehicles if registration data is used for calculation (Baidya and Borken-Kleefeld, 2009; Guttikunda and Mohan, 2014). Unlike passenger cars, commercial vehicles have to register annually and the drawback of overestimation of freight vehicles is unexpected. Indeed it is observed that, in various Indian studies either the domestic sales data compiled by Society of Indian Automobile Manufacturers (SIAM) (Baidya and Borken-Kleefeld, 2009; Guttikunda and Mohan, 2014; Pandey and Venkataraman, 2014) or the registration data (Gurjar et al., 2004; Guttikunda and Kopakka, 2014; Ramachandra, 2009;) is used to construct rolling fleet for both passenger cars and freight vehicles. However, this is to be noticed that the domestic sales published by SIAM are company level (upstream element) estimates in the sequence of elements in the supply chain (as per telephonic communication SIAM, 25th April 2017). This does not represent the wholesalers’ sales from where the vehicles are actually purchased. Thus, it may be possible that vehicle purchased from the company may still be in stock with wholesalers and still off-road. Table 2 shows the

3.3. Emission calculations For emission calculations, we use bottom-up ASIF (Activity–Share– Intensity–Factor) methodology. ASIF is a well-established methodology to calculate tail pipe emissions (Goel et al., 2016, 2015; Guttikunda and Mohan, 2014; Ramachandra, 2009; Schipper et al., 2000). The 125

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

Table 2 Annual freight vehicles registration vs. domestic sales. Year

2008

2009

2010

2011

2012

2013

2014

Registered vehicles (R) Domestic Salea (D) Difference w.r.t registration ((R-D)/R)

439986 322350 0.27

391002 455225 −0.16

632569 592151 0.06

593896 710749 −0.20

938371 698471 0.26

100779 551343 −4.47

646923 533295 0.18

a

Source: SIAM estimates.

The deterioration factors are differentiated by the age group. The emission factor for a particular age is multiplied by the corresponding deterioration factor for that age group to obtain adjusted emission factor. The emission factors, deterioration factors and the adjusted emission factors adopted for the present study are included in the supplementary material. Further, it is to be noted that several studies assume some fraction of defective or poorly maintained vehicles, usually, known as super-emitters (Pandey and Venkataraman, 2014; Yan et al., 2011). The fraction of super-emitters in the case of heavyduty vehicle and the light-duty vehicle is uncertain, thus, no further assumptions regarding this are used in the analysis.

comparison of officially registered freight vehicles to the domestic sales. Given the uncertainty in choosing the two datasets i.e., registration data or domestic sales, the authors compare the total emission estimates from both datasets. In addition to the ambiguity observed in the usage of vehicle registration or domestic sales datasets, unfortunately, there is little information available on the survival rates for the freight vehicles in India. Usually, the actual number of on-road vehicles is predicted by either assuming some borrowed fraction of the total registered vehicles (Singh et al., 2008) or by assuming survival parameters (Baidya and Borken-Kleefeld, 2009) or by using the official registration data as it is (Gurjar et al., 2004; Ramachandra, 2009). An exception to these, a study conducted for the assessment of motor vehicle characteristics for Delhi (Goel and Guttikunda, 2015) develop survival curves for different vehicle classes using fuel station surveys. The survival curves in this study are obtained from the sample survey. The proportion of heavyduty and light-duty vehicle used for interstate and intrastate operations is obtained from the O & D survey. Based on the results, 70% and 20% of the total in-use heavy-duty vehicles and light-duty vehicles are used for interstate mobility, respectively. This indicates that heavy-duty vehicles are usually preferred for long distance freight transport and light-duty vehicles are usually used for intrastate operations. Furthermore, the survival curves thus, developed for the freight vehicles on National Highways are compared with the available survival curves from other studies. The third uncertainty in emission calculation is associated with the use of emission factors. Emission factors measured in terms of pollutant emitted per kilometre driven or per unit power generated by the engine are quite sensitive to the vehicle type, vehicle weight, operating conditions (acceleration, speed), age, terrain type, fuel type and exhaust after treatment. As discussed earlier, the effect of operating conditions, vehicle weight and terrain type are taken into account by the driving cycles (although, these may not truly represent on-road condition). Besides this, the effect of age of the vehicle has also to be taken into account while calculating total emissions. Two aspects can be related to the age of vehicle: the first effect is related to the introduction of new technologies with time thus, making newer engines to comply with stringent emission standards. To take this effect into account emission factors in this study are used corresponding to the specified limits for BS I, BSII, BSIII, BSIV emission standards for diesel vehicles. 100% of the heavy-duty vehicles interviewed are observed to use diesel. A small proportion of 8% light-duty vehicles interviewed are observed to use non-diesel fuel. However, for analysis complete dieselisation of the freight vehicles is assumed. Vehicle model year is cross-matched with the implementation year of the emission standards to come up with the final emission factors. A factor of 1.4 kWh/km is applied to convert the emission standards for the heavy-duty vehicle to gram/km (3 hp-hr/mile~ 1.4 kwh/km; Faiz et al., 1996). The second effect of age is related to the decrease in engine performance with age. Multiplicative deterioration factors are utilised to capture the wear and tear of the engine, deterioration of emissions systems like catalytic convertor with time. Multiplicative factors indicate the maximum emission that a vehicle can achieve in its useful life and is defined as “the ratio of emissions at the end of useful life to emissions at the lowhour test point” (Environmental Protection Agency, 2011). Thus, multiplicative deterioration factors are borrowed from CPCB (2008).

4. Results This section focuses on analysing the existing freight vehicles characteristics used for “interstate mobility” and the potential impacts of the proposed emission and fuel economy standards for freight vehicles on the overall emissions. The results for freight vehicles are classified based on the two categories (1) heavy-duty vehicles (greater than 3.5 t; include 2-axle, 3-axle and multi-axle trucks) (2) light-duty vehicles (less than 3.5 t; include commercial three-wheelers and light commercial vehicles; exclude passenger cars used for commercial purpose). The vehicle classification adopted is as per the classification used for setting the emission standards for freight vehicles in India. 4.1. Age profile Table 3 shows the age distribution of heavy and light-duty vehicles on National highways. From the OD survey, the model year could be obtained, however, for calculations, the month in which the vehicle was manufactured is required. It is assumed, for calculation purposes, that an equal number of commercial vehicles are bought in each month (Goel et al., 2015). The average age of the vehicles is calculated at end of the particular year ((1+2+3+4+5+6+7+8+9+10+11+12)/12 = 6.5 months). Thus, the age of the vehicle is calculated by the addition of the difference between survey year and vehicle registration year, and the factor of 0.54 (6.5/12=0.54). The age distribution obtained for heavy-duty vehicles includes 45.7% of vehicles in the age group 0–4 years, 45.1% in the age group 4–8 years and 9.2% of vehicles in the age group greater than 8 years. In the case of light-duty vehicles, 71% vehicles are observed in the age group 0–4 years, 25.6% vehicles in the age group 4–8 years and 3.3% of vehicles in the age group greater than 8 years. Moreover, based on Table 3 Distribution of freight vehicles according to age.

126

S.No.

Age of vehicle

Heavy-duty vehicle

Light-duty vehicle

1 2 3 4 5 6 7 8 Total

< = 2 years 2 to 4 years 4 to 6 years 6 to 8 years 8 to 10 years 10 to 12 years 12 to 15 years > 15 year (if any)

17.35% 28.34% 26.30% 18.78% 5.59% 2.05% 1.02% 0.57% 100%

23.53% 47.57% 18.41% 7.16% 2.56% 0.51% 0.26% 0.00% 100%

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

LDV_low estimate_India; Baidya and Borken-Kleefeld, 2009). It is to be noted that survival curve for Delhi is estimated using fuel station surveys (Goel and Guttikunda, 2015); however, the survival curve of India is based on some assumed parameters (Baidya and BorkenKleefeld, 2009). “Low estimate” term used in the survival curve representing India refers to the shorter service life of vehicles due to frequent breakdowns, poor road condition and higher accident frequencies. An interesting contrast is observed between the survival curves derived for the three settings. The probability of survival of vehicles of age greater than 10 years in the case of heavy-duty vehicles for national highways, urban area and India is 2%, 63% and 67%, respectively. This clearly indicates the preference for newer vehicles for long distance transport. Probably, this can be attributed to the poor maintenance of roads, risk of equipment breakdown, associated delays and increased logistic costs for older vehicles. However, lower trip lengths in urban areas make it economically feasible for the older trucks to remain in the city system for a longer period of time. Second important observation is related to the flatness of curve estimated to represent heavy-duty and light-duty fleet for India (HDV_low estimate India, LDV_low estimate_India). The fleet survival rates estimated using assumed parameters of modified Weibull parameters appears to be far away from the survey based observations (HDV-NH, LDV-NH, HDV-urban, LDV-urban).

the year of registration and implementation of emission standards it is found that approximately 84%, 13%, 2%, 1% comply with BSIII, BSII, BSI, below BSI emission standards in case of heavy-duty vehicles. For light-duty vehicles, 93%, 6% and 1% of vehicles comply with BSIII, BSII, BSI emission standards. To estimate on-road interstate mobility fleet, survival curves for heavy-duty and light-duty vehicles are developed from the survey data. Weibull distribution is used to compute the survival rates for heavy and light-duty vehicles. The in-use vehicle proportion (the ratio of the number of vehicles observed in particular age group to the number of vehicles observed for the reference year) for various model years is obtained from the sample surveys. These proportions are further applied to the domestic sales/registration data of the corresponding model years to arrive at the in-use vehicle fleet for the year y. The parameters for the survival distribution are thus, estimated by minimising the square of the error between observed in-use vehicles of each model year and in-use vehicles obtained from survival distribution (Eq. (4)). y −1

Min ∑ (Ny − x, observed − Ny − x, survival )2 x

(4)

where Ny−x,observed refers to the in-use vehicles observed in age group y−x and Ny−x, survival refers to the in-use vehicles observed in the age group after applying survival distribution. The estimated parameters for the survival function do not differ significantly if either domestic sales or registration data is used for the calculation. This is attributed to the use of the same dataset at the time for minimisation of the error. However, the total number of vehicles obtained after applying the survival rates to the registration or to the domestic sales certainly differs. The survival function for heavy-duty vehicle and light-duty vehicle are given as per Eqs. (5) and (6), respectively.

⎛ ⎛ y − x ⎞5.26⎞ Sk , y − x = Fk ( y−x ) = exp ⎜⎜−⎜ ⎟ ⎟⎟ for heavy − duty vehicle ⎝ ⎝ 7. 77 ⎠ ⎠

(5)

⎛ ⎛ y − x ⎞6.71⎞ Sk , y − x = Fk ( y−x ) = exp ⎜⎜−⎜ ⎟ ⎟⎟ for light − duty vehicle ⎝ ⎝ 4. 58 ⎠ ⎠

(6)

4.2. Annual vehicle mileage As already discussed, annual vehicle mileage represents the activity of the vehicle and is a crucial parameter in emission estimation. Information regarding annual mileage for freight vehicles in India is scarce as hardly any national level travel inventory is conducted periodically. Uncertainty associated with the use of annual mileage of freight vehicles can be observed in various emission related Indian studies (Goel and Guttikunda, 2015; Gurjar et al., 2004; Ramachandra, 2009). The values of annual mileage for freight vehicles are usually borrowed from the official reports (CPCB, 2005; MoSRTH, 2007). Secondly, issue related to simplifications based on the independence of annual mileage with age is observed. However, this is an observed phenomenon in literature that newer vehicles tend to travel more than the older vehicles (Van Wee et al., 2000; Zachariadis et al., 2001). The reason is associated with the frequent vehicle breakdown, lower reliability, reduced vehicle performance and increased cost per km in the case of older vehicles. Emission calculations based on this independent relationship may result in unreliable estimates. However, the assumption of approximately equal kilometres travelled may hold true only if the operator utilise the vehicle irrespective of age under the given commercial environment (Tsai and Su, 2004). To obtain annual mileage odometer reading observed in surveys is divided by the age of the vehicle. Special attention is paid to not to include the dead odometer readings. Thus, approximately 2.5% entries are excluded from the total samples for the analysis. Figs. 3 and 4 shows the variation of the annual mileage of heavyduty vehicles and light-duty vehicles with age, respectively. The error bars represent the upper and lower bounds at 95% confidence level. The average annual mileage of 51,000 km and 30,000 km are observed for heavy and light-duty vehicles, respectively. Comparing with other studies, Ramachandra (2009) uses the annual travel of 25,000 to 90,000 km for heavy-duty vehicles and 63,000 km for light-duty vehicles for emission calculations. CPCB (2005) suggests annual mileage of 30,000 and 40,000 for heavy-duty and light-duty vehicles, respectively. Pandey and Venkataraman (2014) use annual mileage of 47,000 km for heavy-duty vehicle and 30,000 km for light duty vehicles, respectively. Thus, variation in the use of annual mileage for freight vehicles is clearly indicated. Fig. 5 shows the international comparison of annual mileage for heavy-duty vehicles (Ntziachristos et al., 2008; Tang et al., 2011; US Department of Transportation, 2016). It is observed that the present annual mileage for heavy-duty

Finally, if domestic sales data is used, 1.29 million interstate heavyduty vehicles and 0.36 million interstate light-duty vehicles are estimated to be on-road for the year 2016. On the other hand, if registration data is used, 1.1 million interstate heavy-duty vehicles and 0.35 million interstate light-duty vehicles are estimated to be on-road for the year 2016. Fig. 2 shows the survival rates for light-duty and heavy-duty vehicles used for interstate mobility (HDV-NH, LDV-NH) and its comparison with the survival curves estimated for Delhi (HDV-urban, LDV-urban; estimation based on Supplementary material of Goel and Guttikunda (2015)) and whole of India (HDV_low estimate India,

Fig. 2. Comparison of survival rates estimated for light-duty and heavy-duty vehicles on National Highways with the survival rates derived by Goel and Guttikunda (2015) and Baidya and Borken-Kleefeld (2009).

127

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

Fig. 3. Variation of heavy-duty vehicle annual mileage with age.

vehicles in India is comparable to the annual kilometres driven in the European countries in 2008 and it is far behind the U.S. annual mileage numbers. Annual mileage for heavy-duty vehicles decreases by 30%, 55% and 80% after the age of 5, 10 and 15 years, respectively. The decrease of annual mileage with age in case of light-duty vehicles do not appear to be predominant. This may be either attributed to the low number of samples (approximately 393) being recorded for light-duty vehicles to observe a significant trend or either to the maximum utilisation of vehicle by the operator irrespective of the age of vehicle due to low annual mileage. Fig. 5. Comparison of the reported annual mileage of heavy-duty vehicles by various countries.

4.3. Fuel economy The calculation for greenhouse gas emissions relies on the fuel economy of the vehicle. It is important to know the real world fuel economy values as compared to the numbers declared by the manufacturers. The manufacturer based fuel economy numbers are based on testing procedures under control settings. Using manufacturer based fuel economy numbers may actually lead to an underestimate in the total amount of fuel consumptions and greenhouse gas emissions. Nevertheless, on-road based fuel economy standards are hardly available. Responses regarding the expected/observed fuel economy are required to be captured from the drivers. Yet, these estimates may still not be able to clearly capture the effect of payload, speed, on-board equipment on fuel economy. Table 4 shows the observed fuel economy (km/litre) with the 95% confidence interval. The fuel economy for diesel-run light-duty vehicles ranges in between 11 km/l and 13 km/ litre, whereas for heavy-duty vehicles it is in the range of 3–4 km/l. IISD (2013) reports the fuel economy of 3.5–4 km/l for the truck of capacity greater than 15 t. Also, for light-duty vehicles, the value of 8 km/l is suggested. Malik et al. (2015) report fuel economy of 4.99 km/l for heavy-duty vehicles and 12.08 km/l for light-duty vehicles. In the case of heavy-duty vehicles, no significant trend of fuel economy with age is observed. Rather, the average fuel economy value

Table 4 Fuel economy of freight vehicles vs. age. Age

Heavy-duty vehicle

Light-duty vehicle

< = 2 years 2 to 4 years 4 to 6 years 6 to 8 years 8 to 10 years 10 to 12 years 12 to 15 years > 15 year (if any)

3.44 ± 0.13 3.69 ± 0.10 3.58 ± 0.09 3.57 ± 0.10 3.49 ± 0.18 3.39 ± 0.28 3.66 ± 0.48 4.20 ± 0.70

13.64 ± 0.75 13.78 ± 0.48 13.53 ± 0.81 12.11 ± 1.26 11.30 ± 1.17 – – –

for 15-year-old vehicles is recorded to be higher. Two reasons may be attributed to the observed trend: (a) The observed fuel economy values are strongly affected by the driver's perception, (b) There has been a modal shift with time towards the use of heavier multi-axle vehicles, thus, overall lowering of the fuel economy is observed in the case of new vehicles. In the urban setting (Supplementary material, Goel and Guttikunda, 2015) fuel economy for diesel-run light-duty vehicles ranges in between 4 km/l and 6.6 km/l, whereas for heavy-duty

Fig. 4. Variation of light-duty vehicle annual mileage with age.

128

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

Table 5 Summary of data elements used in ASIF approach for the base year, 2016. Data

Source

Value

Number of interstate heavy-duty vehicles-Domestic Sales Number of interstate light-duty vehicles-Domestic sales Annual mileage heavy-duty vehicles (in km) Annual mileage light-duty vehicles (in km) Emission factors Deterioration factors Fuel economy heavy-duty vehicle (km/l) Fuel economy light-duty vehicle (km/l)

Survival rates heavy duty vehicle (Survey) Survival rates light duty vehicle (Survey) Linear regression model (Survey) Average (Survey) Emission standards India CPCB (2008) Average(Survey) Average(Survey)

1.29 million 0.36 million −2830.1(age)+65,816 30,000 Supplementary Material Supplementary Material 3.58 13.48

vehicles it is in the range of 2–3.85 km/litre. The numbers thus, clearly indicate the variation of acceleration and deceleration patterns, congestion level on national highways and urban roads.

• • •

4.4. Total emissions base scenario-2016 The parameters required for ASIF approach are collected by using the survey-based methodology and secondary sources. Table 5 summarises the source of each of these elements and the corresponding values used for the calculation of total emissions for the base year. Both, linear and logarithmic regression lines are used to test the relationship of age of heavy-duty vehicle with annual mileage. The linear regression line (equation in Table 5) with highest R2 of 0.83 is chosen to explain the variance in the observed data. Further, as no predominant trend is observed between age and annual mileage for light-duty vehicles, single average annual mileage is used for all these vehicles. The base scenario uses BSIII standards until 2016 instead of the implementation of BSIV standards in few selected cities in 2010. This is motivated by the fact that staggered implementation of emission standards end up in negating the expected benefits of new emission standards due to the availability of low-quality fuel in different states (Guttikunda and Mohan, 2014). Finally, the calculations are performed as per Eq. (1) to calculate the total emissions for criteria pollutants. The results are discussed for NOx and PM as these are the main pollutants associated with diesel operated freight vehicles. A total of 545 Gg of NOx and 15.36Gg of PM are contributed by interstate heavy-duty vehicles in the base year 2016. Interestingly, the pollution produced by the vehicles which are greater than six years old is 10% and 12% of the total NOx and PM emissions, respectively. In the case of light-duty vehicles, a total of 8.4 Gg of NOx and 1.24Gg of PM are produced. Further, effects of the following cases are examined on the total emissions, and, on the emissions within each age group with respect to the reference scenario. The results presented are in reference to heavyduty vehicles.

• • • • •

• • • •

vehicles. R+avg – If registration data is used to calculate on-road vehicles and observed average annual mileage of 51,000 km is used in case of heavy-duty vehicles for all age groups. R+lit- If registration data is used to calculate on-road vehicles and annual mileage used is equal to 57,500 km. R+cpcb – If registration data is used to calculate on-road vehicles and annual mileage of 40,000 as reported by CPCB (2005) is used for all vehicles P+ line –If the proportion of 60% is applied to total cumulative heavy-duty vehicles (Baidya and Borken-Kleefeld, 2009) to obtain on-road vehicles and annual mileage varies with age as per linear regression equation. Use of proportion results in 1.8 million interstate heavy-duty vehicles. P+ avg- If the proportion of 60% is applied to total cumulative heavy- duty vehicles to obtain on-road vehicles and observed average annual mileage of 51,000 km is used in case of heavy-duty vehicles for all age groups. P+ lit- If the proportion of 60% is applied to total cumulative heavyduty vehicles to obtain the on-road vehicles and annual mileage used is equal to 57,500 km. P+cpcb- If the proportion of 60% is applied to total cumulative heavy-duty vehicles to obtain the on-road vehicles and annual mileage of 40,000 as reported by CPCB (2005) is used for all vehicles

Fig. 6 shows the underestimation or overestimation of total NOx and PM emissions as a result of the use of different approaches with respect to the base case. It is observed that using registration data for heavy-duty vehicles instead of domestic sales would have underestimated NOx and PM in each of the cases with respect to the base case scenario. Opposite to this is the trend of overestimation of emissions if the proportion was used to calculate on-road vehicles. Indeed, overestimation of 60% is observed with respect to the base scenario if the proportion and annual mileage varying with age is used to predict total emissions. Further, underestimation of emissions by 8% and 6% for

S+line – (Base scenario). If domestic sales data is used to calculate on-road vehicles and annual mileage varies with age as per linear regression equation. S+avg – If domestic sales data is used to calculate on-road vehicles and observed average annual mileage of 51,000 km is used in case of heavy-duty vehicles for all age groups. S+lit – If domestic sales data is used to calculate on-road vehicles and the annual mileage used is the mean of 25,000 km and 95,000 km as reported by Ramachandra (2009). Thus, annual mileage of 57,500 km is used for all age groups. S+cpcb - If domestic sales data is used to calculate on-road vehicles and the annual mileage of 40,000 as reported by CPCB (2005) is used for all vehicles. R+line - If registration data is used to calculate on-road vehicles and annual mileage varies with age as per linear regression equation. Use of registration data results in 1.1 million interstate heavy-duty

Fig. 6. Estimated percentage change in total NOx and PM emissions with respect to the base case (s+line).

129

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

Table 6 Estimated percentage change in NOx emissions by different age groups w.r.t base case (s+line). Age

s+avg

s+lit

s+cpcb

r+line

r+avg

r+lit

r+cpcb

p+line

p+avg

p+lit

p+cpcb

< =2 2 to 4 4 to 6 6 to 8 8 to 10

−19% −9% 1% 14% 28%

−9% 3% 14% 28% 44%

−37% −28% −20% −11% 0.3%

4% −61% −6% −3% −5%

−16% −64% −5% 10% 21%

−5% −60% 7% 24% 37%

−34% −72% −25% −14% −5%

56% 64% 64% 64% 64%

26% 49% 66% 86% 109%

42% 68% 87% 110% 136%

−1% 17% 30% 46% 64%

NOx and PM, respectively, is observed if a uniform annual average of 51,000 km is used as compared to the base case. Other interesting insights are obtained when the comparisons of total emissions for NOx and PM are carried within the age group as a result of the use of different approaches. It is observed that if a uniform annual mileage of 51,000 km and 57,500 km is used for all vehicles, the overestimation in the total NOx emissions in case of 8 years older vehicles is approximately 28% and 44%, respectively (Table 6). However, the effect of the overestimation (0.3%) of emissions for the older vehicles is diluted due to lower annual mileage used by CPCB (2005). Yet, at the same time underestimation of 20–37% is observed in the case of new vehicles. Use of registration data with vehicle age dependent annual mileage shows 5% underestimation with reference to the base case. Use of the proportion for the calculation of the on-road vehicle is associated with the highest percentage of overestimation for the older vehicles.

Fig. 8. Estimated heavy-duty vehicles emission outlook for PM from 2016 to 2026.

4.5. Emission outlook This section presents the emission outlook for the current policy interventions by the Indian Government to reduce pollution from the transport sector. Following scenarios are tested: a) Business as Usual (BAU): This is the base scenario where no improvements are observed in current emission standards. BS III standards are used for the calculation of the base scenario. b) B417: Introduction of BS IV emission standards nationwide in 2017. c) B620: Introduction of BSIV emission standards nationwide in 2017 followed by the implementation of BSVI emission standards in 2020.

Fig. 9. Estimated heavy-duty vehicles emission outlook for CO from 2016 to 2026.

be observed. PM emissions show a significant reduction of 80% through the introduction of BS IV standards nationwide. A total reduction of 89% will be achieved in PM emissions if BSVI standards are further implemented. In the context of CO emissions, 29% reduction in emissions is expected if BSIV emission standards are implemented. Further, implementation of BS VI emission does not affect the emission trend due to the similar CO emission limits for BSIV and BSVI. Similarly, in the case of light-duty vehicles, 50% reduction in NOx will be observed by 2025 if BS IV emission standards are implemented in 2017. If BSIV emission standards are followed by BSVI emission standards in 2020, 84% reduction in NOx emissions will be observed. Further, PM emissions show a reduction of 50% through the introduction of BS IV standards nationwide. A total reduction of 91% will be achieved in PM emissions if BSVI standards are further implemented. In the context of CO emissions, 22% reduction in emissions is expected if BSIV emission standards are implemented.

BAU uses no improvements in BSIII standards. Further, based on the average sales trend we assume year on year growth of 5% and 10% for heavy-duty vehicles and light-duty vehicles respectively, in India. Constant survival rates are used to estimate the on-road fleet for subsequent years for emission calculations. Figs. 7–9 show the projected emissions for pollutants NOx, PM and CO in the case of heavy-duty vehicles, respectively. 30% reduction in NOx will be observed by 2026 if BS IV emission standards are implemented in 2017. If BSIV emission standards are followed by BSVI emission standards in 2020, 89% reduction in NOx emissions will

4.6. Fuel and CO2 projections The fixing of the time-limit for the achievement of the fuel consumption targets acts as an important starting point for negotiations between government and industry (Plotkin, 2009). Perhaps, the setting of achievable targets and their time frame play an equally important role in determining the success of the fuel economy policies. In literature different increment (fuel economy km/l)/decrement (fuel consumption, l/km) rates are used to develop fuel and CO2 projections. Huo et al. (2007) develop three fuel economy scenarios based on U.S.

Fig. 7. Estimated heavy-duty vehicles emission outlook for NOx from 2016 to 2026.

130

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

Table 7 Comparison of fuel economy scenarios for heavy-duty vehicles. Year

Fuel economy improvement (km/l)

BAU fuel consumption (mn tonne)

FEC fuel consumption (mn tonne)

% change in fuel consumption w.r.t baseline

BAU CO2 (million tonne)

FEC CO2 (million tonne)

% change in CO2 emissions w.r.t baseline

2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026

3.58 3.72 3.87 4.01 4.15 4.30 4.48 4.66 4.83 5.01 5.19

16.68 16.39 16.63 17.20 18.09 19.15 20.22 21.27 22.33 23.45 24.62

16.68 15.76 15.40 15.35 15.60 15.96 16.18 16.36 16.54 16.75 16.98

0.00 −3.85 −7.41 −10.71 −13.79 −16.67 −20.00 −23.08 −25.93 −28.57 −31.03

52.90 51.98 52.72 54.53 57.37 60.72 64.11 67.43 70.82 74.36 78.08

52.90 49.98 48.81 48.69 49.46 50.60 51.29 51.87 52.46 53.11 53.85

0.00 −3.85 −7.41 −10.71 −13.79 −16.67 −20.00 −23.08 −25.93 −28.57 −31.03

l.

National Academy of Science Paths for light-duty vehicles and use Japanese heavy-duty fuel consumption targets as a reference for Chinese heavy-duty vehicles. The paths represent the rate at which different technologies investments can be made to achieve a particular fuel economy target. The important inputs required for the projections are—the stock turnover over years differentiated by age group, fuel economy by age group and annual mileage by age group. Further, “rebound” effect due to increase in annual mileage of new vehicles as a result of improved fuel economy is not considered in the analysis. Karplus et al. (2013) use two policy scenarios w.r.t the year 2010 (a) Sharp policy: where fuel consumption is reduced to half by 2025 (37.5% reduction in next 15 years) and remain constant thereafter up till 2050; (b) Gradual policy: where steady incremental reductions are done annually to meet similar targets by 2050 (27.2% reduction in next 15 years). The authors use single targets that are to be complied by all new vehicles. However, the use of single fuel economy targets limits the scope of analysis by not setting differential fuel economy targets for different types of manufacturers in the market. ICCT (2016) examines the potential impacts on the fuel consumption of rigid truck and tractor and trailor fleet of India with the adoption of presently commercialised or demonstrated fuel efficiency technologies. The study considers three scenarios for tractor-trailers, according to which fuel consumption reduces by 52% by 2040 (Increment; decrement rate 2.9%), by 2035 (moderate; decrement rate 3.6%) and by 2030 (accelerated; decrement rate 4.8%) with respect to 2015. Similarly, decrement rates of 1.8%, 2.2% and 2.9% are assumed for the rigid truck. Furthermore, Plotkin (2009) after careful consideration and review of various studies suggests that 30–50% improvement in fuel economy over 12–15 years may be used as a reasonable starting point for light-duty vehicles. Taking into account the above factors, two separate scenarios related to the fuel economy standards are tested to project the GHG emissions and fuel consumption.

There are several limitations related to these projections. First, setting a single fuel economy target is a naïve approach and does not differentiate between different manufacturers as discussed earlier in the Section 2.2. Second, we use more aggregate classification for the freight vehicles i.e., less than 3.5 t or more. We do not differentiate between fuel economies of a 2-axle truck from a 3-axle or multi-axle truck. They are all classified under the heavy-duty category. Thirdly, the future sales projections for the different type of vehicles are not considered. It may be possible as seen in the case of U.S. that market may shift to the heavier weight class, thus, reducing the fuel economy. Also, consumer vehicle choices and behaviour after the implementation of the fuel economy standards are not considered in these projections. Moreover, the increase in the annual mileage of vehicles with improved fuel economy is not modelled. Based on the calculation, if fuel economy standards are phased in, the CO2 emissions and fuel consumption of heavy-duty vehicles are expected to reduce by 31% in 2026 (Table 7). Further, the phasing in of fuel economy standards for light-duty vehicles will lead to the reduction of GHG emissions and fuel consumption by 26% in 2026. 5. Conclusions and future work Given the nationwide concern towards the growing energy consumption and emissions from the freight vehicles, the present work aims at analysing the potential impacts of proposed fuel emission and economy standards in India for interstate freight fleet. The study contributes to the literature majorly in three ways:

•

Previous studies use different assumptions for survival rates, annual mileage or fuel economy to calculate the total contribution of freight vehicles to overall pollution. This study attempts to narrow this gap by developing an emission inventory for freight vehicles used for interstate mobility. The characteristics in terms of age distribution, the variation of annual mileage and fuel economy with age are examined. Implication of various assumptions on the total amount of emissions is tested with respect to the base scenario developed using surveybased results.

a) Business as Usual: No improvements in current fuel economy standards. b) Accelerated improvements in the fuel economy standards by 20% by 2021 and 45% by 2026 for heavy-duty vehicles and improvements in the fuel economy standards by 15% by 2021 and 35% by 2026 for light-duty vehicles.

• •

Two reasons may explain the choosing of accelerated scenario for the fuel economy projections: First, the fleet turnover in the case of freight vehicles seems to be high. Thus, much can be achieved if the fuel efficiency technologies are introduced at an earlier stage. Second, the accelerated scenario presents the best achievable goals with the presently available fuel efficiency technologies. For the calculation of CO2 emissions and fuel consumption, we use single fleet average fuel economy for heavy and light-duty vehicles as no particular trend of fuel economy with age is observed. The carbon content in diesel fuel is assumed to be 86%. Further, the density of the fuel is taken as 0.82 kg/

The objectives of the study are achieved by conducting O & D surveys and traffic volume counts at ten locations on national highways to capture the fleet characteristics. The results indicate the presence of newer fleet for long distance transport as compared with the urban area of Delhi. 46% and 71% of the fleet lies in the age group of 0–4 years for heavy and light-duty vehicles respectively. Only, 3.64% and 0.77% of heavy-duty vehicle and light-duty vehicle are greater than 10 years old, respectively. Comparison between the three survival curves highlights significant variations between survival probabilities thus, calling for the need of country representative survival estimates. The average annual 131

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

for planned freight policies for the betterment of environment and society as a whole.

mileage of 51,000 km and 30,000 km are observed for heavy and lightduty vehicles, respectively. Uncertainties associated with the use of annual mileage of freight vehicles for emission calculations are also discussed. For heavy-duty vehicles, the decreasing relationship of annual mileage with age is captured. Annual mileage for heavy-duty vehicles decreases by 30%, 55% and 80% after the age of 5, 10 and 15 years, respectively. However, in the case of light-duty vehicles, a similar relationship between annual mileage and age is not found to be predominant. The fuel economy for diesel-run light-duty vehicles ranges in between 11 km/l and 13 km/l, whereas, for heavy-duty vehicles, it is in the range of 3–4 km/l. The results for the urban setting show the fuel economy to range from 4 to 6.6 km/l for lightduty vehicles and 2 km/l to 3.85 km/l for heavy-duty vehicles, thus, representing the operating conditions in the two cases. The lower fuel economy of new heavy-duty vehicles in comparison to old vehicles is not clearly understood and is supposed to be affected by drivers’ perception or modal shift with time towards the heavier fleet. Further, comparison of various assumptions (annual mileage and on-road vehicles) used for calculating emission contribution by the freight vehicles is examined. A significant overestimation in emissions is observed if a particular proportion is applied to the cumulative registered vehicles to calculate on-road vehicles. Indeed, this approach is also associated with significant overestimation of the emissions by old vehicles. The analysis for three emission scenarios indicates a clear gain in early transitioning to the stringent emission standards. The emission reduction varies with the different type of pollutants. For NOx, 89% reduction is expected if BSIV standards are followed by BSVI standards in 2020 for heavy-duty vehicles. However, in the case of PM, a major gain is achieved by the implementation of BSIV standards in 2017 (80% reduction) for heavy-duty vehicles. In the case of light-duty vehicles, 84% reduction in NOx emissions will be observed by the implementation of BSVI standards in 2020. A total reduction of 91% will be achieved in PM emissions if BSVI standards are further implemented for light-duty vehicles. Finally, accelerated phasing of fuel economy standards is expected to reduce CO2 emissions and fuel consumption by 31% and 26% for heavy-duty and light-duty vehicle in 2026, respectively. Although the scope of the current study is restricted to the national highways, the work is to be further extended to the different type of urban areas and different type of roads. For instance, the characteristics related to a metropolitan city may be quite different from a smallmedium sized city. Averaging the results for different areas will thus help to predict a representative survival curve for the national fleet of heavy and light-duty vehicles. This may also help us to predict the potential impacts of the proposed scrappage policy in India. Given the available data set, it may be too early to comment on this area. Further, the emission factors (g/km) based on which emission calculations for heavy-duty and light-duty vehicles are performed, do not correspond to the real world driving and payload conditions. Comprehensive documentation of on-road variation of emissions will help to reduce the uncertainties in the results. Thirdly, the proportion of super-emitters among the freight vehicles is required to be investigated to further improve emission estimates. Fourthly, future projections for the fuel economy standards need to be further improved. Long-term projections will require a complete understanding of presently available fuel efficiency technology, the timeframe of their proliferation, market and societal response to fuel efficiency standards. Establishing disaggregate baseline scenarios for the different class of freight vehicles should be taken into account seriously. However, from the aggregate analysis performed in the study, it is clear that much of the benefits in emission reduction can be achieved if emission standards are nationwide and timely implemented. The higher turnover rate of freight fleet offers a great scope of penetration of new fuel efficient technologies in the system. This is the need of the hour that we appropriately utilise the potential lying ahead, and push

Acknowledgements The research is partially funded by Shakti Sustainable Energy Foundation (G15 SSEF-178), New Delhi, Delhi, India. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.enpol.2017.05.053. References An, F., Earley, R., Green-Weiskel, L., 2011. Global overview on fuel efficiency and motor vehicle emission standards: policy options and perspectives for international cooperation. U. N. Backgr. Pap. 3. An, F., Sauer, A., 2004. Comparison of passenger vehicle fuel economy and greenhouse gas emission standards around the world. Pew Cent. Glob. Clim. Change 25. Apelbaum, J., 2009. A case study in data audit and modelling methodology—Australia. Energy Policy 37, 3714–3732. Atabani, A.E., Badruddin, I.A., Mekhilef, S., Silitonga, A.S., 2011. A review on global fuel economy standards, labels and technologies in the transportation sector. Renew. Sustain. Energy Rev. 15, 4586–4610. Auto Fuel Vision and Policy 2025 Report, 2014. Government of India,New Delhi. Baidya, S., Borken-Kleefeld, J., 2009. Atmospheric emissions from road transportation in India. Energy Policy 37, 3812–3822. Bandivadekar, A., 2015. Direct progression from Bharat IV to Bharat VI national vehicle emission standards in India, ICCT [WWW Document]. URL 〈http://www.theicct. org/leapfrogging-bsIV-to-bsVI-india〉. Bansal, G., Bandivadekar, A., 2013. Overview of India’s Vehicle Emissions Control Program: Past Successes And Future Prospects, ICCT. [WWW Document]. URL 〈http://www.theicct.org/sites/default/files/publications/ICCT_IndiaRetrospective_ 2013.pdf〉. CAI-Asia, 2011. Best Practices in Green Freight for an Environmentally Sustainable Road Freight Sector in Asia [WWW Document]. URL 〈http://www.uncrd.or.jp/content/ documents/6EST-5A-BGP.pdf〉. Census of India [WWW Document], 2011. Gov. India. URL 〈http://censusindia.gov.in/〉 (accessed 21 December 2016). Cheah, L., Heywood, J., 2011. Meeting US passenger vehicle fuel economy standards in 2016 and beyond. Energy Policy 39, 454–466. Clark, M.L., Jessup, E.L., Casavant, K., 2002a. Freight Truck Origin and Destination Study: Methods, Procedures, and Data Dictionary. Washington State University, Washington, United States. Clark, N.N., Kern, J.M., Atkinson, C.M., Nine, R.D., 2002b. Factors affecting heavy-duty diesel vehicle emissions. J. Air Waste Manag. Assoc. 52, 84–94. CPCB, 2008. Emission Factor development for Indian Vehicles [WWW Document]. URL 〈http://www.cpcb.nic.in/Emission_Factors_Vehicles.pdf〉. CPCB, 2005. Transport Fuel Quality of the year 2005. Dhar, S., Shukla, P.R., 2015. Low carbon scenarios for transport in India: co-benefits analysis. Energy Policy 81, 186–198. DieselNet, 2017. EU: Cars: Greenhouse Gas Emissions [WWW Document]. URL 〈https://www.dieselnet.com/standards/eu/ghg.php〉. DOT, U.S., 2014. Highway performance monitoring system field manual. US Dep. Transp. Fed. Highw. Adm. Environmental Protection Agency, 2011. 40 CFR 1033.901 - Definitions [WWW Document]. URL 〈https://www.law.cornell.edu/cfr/text/40/1033.901〉. Faiz, A., Weaver, C.S., Walsh, M.P., 1996. Air Pollution from Motor Vehicles: Standards and Technologies for Controlling Emissions. World Bank Publications, Washington, DC, United States. Goel, R., Guttikunda, S.K., 2015. Evolution of on-road vehicle exhaust emissions in Delhi. Atmos. Environ. 105, 78–90. Goel, R., Guttikunda, S.K., Mohan, D., Tiwari, G., 2015. Benchmarking vehicle and passenger travel characteristics in Delhi for on-road emissions analysis. Travel Behav. Soc. 2, 88–101. Goel, R., Mohan, D., Guttikunda, S.K., Tiwari, G., 2016. Assessment of motor vehicle use characteristics in three Indian cities. Transp. Res. Part D Transp. Environ. 44, 254–265. Gurjar, B.R., Van Aardenne, J.A., Lelieveld, J., Mohan, M., 2004. Emission estimates and trends (1990–2000) for megacity Delhi and implications. Atmos. Environ. 38, 5663–5681. Guttikunda, S.K., Kopakka, R.V., 2014. Source emissions and health impacts of urban air pollution in Hyderabad, India. Air Qual. Atmos. Heal. 7, 195–207. Guttikunda, S.K., Mohan, D., 2014. Re-fueling road transport for better air quality in India. Energy Policy 68, 556–561. Hill, N., Finnegan, S., Norris, J., Brannigan, C., Wynn, D., Baker, H., Skinner, I., 2011. Reduction and testing of greenhouse gas (GHG) emissions from heavy duty vehicles– Lot 1: strategy. Final Rep. to Eur. Comm. Clim. Action. AEA Technol. plc. Huai, T., Shah, S.D., Miller, J.W., Younglove, T., Chernich, D.J., Ayala, A., 2006. Analysis of heavy-duty diesel truck activity and emissions data. Atmos. Environ. 40,

132

Energy Policy 108 (2017) 121–133

L. Malik, G. Tiwari

Prozzi, J., Wong, C., Harrison, R., 2004. Texas truck data collection guidebook. Pucher, J., Korattyswaropam, N., Mittal, N., Ittyerah, N., 2005. Urban transport crisis in India. Transp. Policy 12, 185–198. Ramachandra, T.V., 2009. Emissions from India's transport sector: statewise synthesis. Atmos. Environ. 43, 5510–5517. RITES, 2009. Total Transport System System Study on Traffic Flows and Modal Costs (Highways, Railways, Airways and Coastal Shipping) [WWW Document]. URL 〈http://planningcommission.nic.in/reports/genrep/index.php?Repts=rites.html〉. Schipper, L., 2011. Automobile use, fuel economy and CO 2 emissions in industrialized countries: encouraging trends through 2008? Transp. Policy 18, 358–372. Schipper, L., Leather, J., Fabian, H., 2009. Transport and carbon dioxide emissions: forecasts, options analysis, and evaluation [WWW Document]. URL 〈http://www. indiaenvironmentportal.org.in/files/Transport-CO2-Emissions.pdf〉. Schipper, L., Marie-Lilliu, C., Gorham, R., 2000. Flexing the Link between Transport and Greenhouse Gas Emissions-A Path for the World Bank. Singh, A., Gangopadhyay, S., Nanda, P.K., Bhattacharya, S., Sharma, C., Bhan, C., 2008. Trends of greenhouse gas emissions from the road transport sector in India. Sci. Total Environ. 390, 124–131. Tang, D., Ding, Y., Yin, H., Wagner, D.V., 2011. Developing a firstever national mobile source emissions inventory for China. Balt. Maryl. TERI, 2015. Air Pollution and Health [WWW Document]. URL 〈http://www.teriin.org/ projects/teddy/pdf/air-pollution-health-discussion-paper.pdf〉 (accessed 25 January 2017). Tsai, M.-C., Su, C.-C., 2004. Scenario analysis of freight vehicle accident risks in Taiwan. Accid. Anal. Prev. 36, 683–690. U.S. Department of Transportation, 2016. Table 1-35: U.S. Vehicle-Miles (Millions) [WWW Document]. Natl. Transp. Stat. URL 〈https://www.rita.dot.gov/bts/sites/ rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_ 01_35.html〉 (accessed 1 February 2017). Van Wee, B., Moll, H.C., Dirks, J., 2000. Environmental impact of scrapping old cars. Transp. Res. Part D Transp. Environ. 5, 137–143. WHO, 2016. WHO’s Ambient Air Pollution database. [WWW Document]. URL 〈http:// www.who.int/en/〉 (accessed 12 January 2016). Williams, T.A., Chigoy, B., Borowiec, J., Glover, B., 2016. Methodologies Used to Estimate and Forecast Vehicle Miles Traveled (VMT). Yan, F., Winijkul, E., Jung, S., Bond, T.C., Streets, D.G., 2011. Global emission projections of particulate matter (PM): i. Exhaust emissions from on-road vehicles. Atmos. Environ. 45, 4830–4844. Zachariadis, T., 2006. On the baseline evolution of automobile fuel economy in Europe. Energy Policy 34, 1773–1785. Zachariadis, T., Ntziachristos, L., Samaras, Z., 2001. The effect of age and technological change on motor vehicle emissions. Transp. Res. Part D Transp. Environ. 6, 221–227. Zamboni, G., André, M., Roveda, A., Capobianco, M., 2015. Experimental evaluation of Heavy Duty Vehicle speed patterns in urban and port areas and estimation of their fuel consumption and exhaust emissions. Transp. Res. Part D Transp. Environ. 35, 1–10. Zamboni, G., Malfettani, S., André, M., Carraro, C., Marelli, S., Capobianco, M., 2013. Assessment of heavy-duty vehicle activities, fuel consumption and exhaust emissions in port areas. Appl. Energy 111, 921–929.

2333–2344. Huo, H., Wang, M., Johnson, L., He, D., 2007. Projection of Chinese motor vehicle growth, oil demand, and CO2 emissions through 2050. Transp. Res. Rec. J. Transp. Res. Board, 69–77. ICCT, 2016. Estimating the fuel efficiency technology potential of heavy-duty trucks in major markets around the world.URL 〈https://www.globalfueleconomy.org/media/ 404893/gfei-wp14.pdf〉. IEA, 2015. India OutlookEnergy [WWW Document]. International Energy Agency. URL 〈https://www.iea.org/publications/freepublications/publication/ IndiaEnergyOutlook_WEO2015.pdf〉 (accessed 20 January 2017). International Energy Agency (IEA), 2009. Transport, Energy and CO2: Moving Towards Sustainability [WWW Document]. International Energy Agency. URL 〈https://www. iea.org/publications/freepublications/publication/transport2009.pdf〉 (accessed 20 January 2017). International Institute for Sustainable Development (IISD), 2013. The Impacts of India’s Diesel Price Reforms on the Trucking Industry. The International Institute for Sustainable Development. [WWW Document]. URL 〈https://www.iisd.org/gsi/sites/ default/files/ffs_india_irade_trucking.pdf〉 (accessed 25 December 2016). Karplus, V.J., Paltsev, S., Babiker, M., Reilly, J.M., 2013. Should a vehicle fuel economy standard be combined with an economy-wide greenhouse gas emissions constraint? Implications for energy and climate policy in the United States. Energy Econ. 36, 322–333. Kumapley, R., Fricker, J., 1996. Review of methods for estimating vehicle miles traveled. Transp. Res. Rec. J. Transp. Res. Board, 59–66. Lee, E.T., Wang, J., 2003. Statistical Methods for Survival Data Analysis. John Wiley & Sons, Hoboken, New Jersey, United States. Malik, L., Tiwari, G., Mohan, D., 2015. Assessment of Heavy-Duty Vehicle Characteristics in Delhi. URL 〈http://staging.unep.org/transport/lowcarbon/PDFs/ heavyDutyVehiclesDelhi.pdf〉. McKinnon, A.C., Piecyk, M.I., 2009. Measurement of CO2 emissions from road freight transport: a review of UK experience. Energy Policy 37, 3733–3742. MoRTH, 2016. Road Transport Year Book (2013–14 and 2014–2015). New Delhi. MoRTH, 2011. Government of India. New Delhi.URL 〈https://data.gov.in/〉 (accessed 20 April 2017). MoSRTH, 2007. The Working Group Report on Road Transport for the Eleventh Five Year Plan, 2007. Planning Commission, New Delhi, India. National Cooperative Highway Research Program, 2011. Freight Transportation Surveys:A Synthesis of Highway Practice. Washington, D.C. Ntziachristos, L., Mellios, G., Kouridis, C., Papageorgiou, T., Theodosopoulou, M., Samaras, Z., Zierock, K.H., Kouvaritakis, N., Panos, E., Karkatsoulis, P., 2008. European database of vehicle stock for the calculation and forecast of pollutant and greenhouse gases emissions with TREMOVE and COPERT. Lab Appl. Thermodyn. 8, 260. Pandey, A., Venkataraman, C., 2014. Estimating emissions from the Indian transport sector with on-road fleet composition and traffic volume. Atmos. Environ. 98, 123–133. PCRA, 2013. Study on Fuel Efficiency of Transport Sector by ICRA. Planning Commission, 2013. Twelfth Five Year Plan (2012-17). Government of India, New Delhi. Plotkin, S.E., 2009. Examining fuel economy and carbon standards for light vehicles. Energy Policy 37, 3843–3853.

133