• Open Access

Implementing Trans-Boundary Infrastructure-Based Greenhouse Gas Accounting for Delhi, India

Data Availability and Methods


Address correspondence to: Abel Chavez, Center for Sustainable Infrastructure Systems (CSIS), University of Colorado Denver, 1201 5th Street, Room AD240, PO Box 173364, Denver, CO 80217-3364, USA. Email: abel.chavez@ucdenver.edu


Community-wide greenhouse gas (GHG) emissions accounting is confounded by the relatively small spatial size of cities compared to nations—due to which, energy use in essential infrastructures serving cities, such as commuter and airline transport, energy supply, water supply, wastewater infrastructures, and others, often occurs outside the boundaries of the cities using them. The trans-boundary infrastructure supply chain footprint (TBIF) GHG emissions accounting method, tested in eight U.S. cities, incorporates supply chain aspects of these trans-boundary infrastructures serving cities, and is akin to an expanded geographic GHG emissions inventory. This article shows the results from applying the TBIF method in the rapidly developing city of Delhi, India.

The objectives of this research are to (1) describe the data availability for implementing the TBIF method within a rapidly industrializing country, using the case of Delhi, India; (2) identify methodological differences in implementation of the TBIF method between Indian versus U.S. cities; and (3) compare broad energy use metrics between Delhi and U.S. cities, demonstrated by Denver, Colorado, USA, whose energy use characteristics and TBIF GHG emissions have previously been shown to be similar to U.S. per capita averages.

This article concludes that most data required to implement the TBIF method in Delhi are readily available, and the methodology could be translated from U.S. to Indian cities. Delhi's 2009 community-wide GHG emissions totaled 40.3 million metric tonnes of carbon dioxide equivalents (t CO2-eq), which are normalized to yield 2.3 t CO2-eq per capita; nationally, India reports its average per capita GHG emissions at 1.5 t CO2-eq. In-boundary GHG emissions contributed to 68% of Delhi's total, where end use (including electricity) energy in residential buildings, commercial and industrial usage, and fuel used in surface transportation contributed 24%, 19%, and 21%, respectively. The remaining 4% of the in-boundary GHG emissions were from waste disposal, water and wastewater treatment, and cattle. Trans-boundary infrastructures were estimated to equal 32% of Delhi's TBIF GHG emissions, with 5% attributed to fuel processing, 3% to air travel, 10% to cement, and 14% to food production outside the city.


Cities are increasingly recognizing their role in global greenhouse gas (GHG) emissions. More than 1,000 cities have signed onto the International Council for Local Environmental Initiatives (ICLEI),1 Cities for Climate Protection (CCP) (a framework for engaging local governments in political climate action commitments [ICLEI 2009 2011]), and the Mexico City Pact (MCP) (an agreement by more than 140 world mayors to establish GHG emissions inventories and mitigation plans [WMSC 2010]). Outcomes from these efforts include public domain items such as the carbonn® Cities Climate Registry (cCCR), a voluntary online tool where cities report on their GHG inventories and mitigation commitments (cCCR 2010). To date, however, these tools, while very valuable, have not incorporated trans-boundary GHG emissions associated with human activities in cities, which have been shown to be quite significant (Hillman and Ramaswami 2010; Kennedy et al. 2009; Ramaswami et al. 2008).

Understanding GHG emissions associated with cities in India, China, and the United States is important due to their contribution to world totals. A report by the International Energy Agency (IEA) notes that India, China, and the United States together constitute 42% of the world's population, and 46% of the world's carbon dioxide (CO2) emissions from fuel combustion (IEA 2010). Moreover, in India, China, and the United States, 30%, 44%, and 82% of people live in urban areas, respectively (World Bank 2010). With rapid urbanization seen especially in Indian and Chinese cities, quantification of GHG emissions associated with cities becomes important. However, GHG emissions accounting for cities is confounded by the relatively small spatial size of cities compared to nations, due to the fact that essential infrastructures—commuter and airline transport, energy supply, water supply, wastewater infrastructures, and others—cross city boundaries; hence, energy use to provide these services often occurs outside the boundaries of the cities using them (Hillman and Ramaswami 2010; Ramaswami et al. 2008). Significant trade of other goods and services also occurs across city boundaries, with associated embodied GHGs.

Two approaches to GHG emissions footprinting (see the review by Ramaswami et al. [2011] and the article by Chavez and Ramaswami [2011]) can be used to alleviate these challenges. The two approaches are the trans-boundary infrastructure supply chain footprint (TBIF) and the consumption-based footprint (CBF).

The TBIF method utilizes the concept of scopes from corporate GHG emissions accounting protocols to include both in-boundary and trans-boundary GHG emissions associated with key community-wide activities; hence it has also been referred to as an expanded geographic inventory or a community-wide infrastructure GHG footprint. The TBIF method recognizes that cities include both producers and consumers, and focuses on infrastructure supply chains that serve the entire community as a whole. The GHG emissions accounted for by the TBIF method are (1) direct in-boundary GHG emissions (scope 1), (2) indirect GHG emissions from the generation of purchased electricity (scope 2), and (3) GHG emissions from essential trans-boundary infrastructures serving cities (scope 3), such as water supply, transportation fuels, airline and commuter travel, and other critical supply chains. The inclusion of trans-boundary infrastructures (scope 3) warrants careful allocation of GHGs to avoid double counting (Ramaswami et al. 2008). For example, infrastructures such as large electric power plants or oil refineries are easily recognized within city boundaries, and their GHGs can be readily allocated based on local demand, thus reducing double counting. The TBIF supports infrastructure planning for the city as a whole, maintaining residential, commercial, and industrial activities together, consistent with the geopolitical definition. The TBIF method captures life cycle GHGs from essential infrastructures serving cities; however, it does not account for life cycle GHGs of other, non-infrastructure goods and services consumed by households, or other non-infrastructure supply chains serving local industries, because such data are often proprietary. Indeed, incorporating key industrial supply chains into the TBIF could enhance this method because the TBIF addresses both consumers and producers in cities. Improved blended metrics that combine GHGs per capita and GHGs per unit productivity may be needed.

The second approach is a CBF, which quantifies the full life cycle GHG emissions from economic final consumption in a city defined as household expenditures, government expenditures, and business capital investments. CBFs have traditionally been conducted at the scale of households, using household consumer expenditure surveys (CESs) (Jones and Kammen 2011), with regional/national production matrices, coupled with sector-specific GHG emissions intensities (e.g., Lenzen and Peters 2010). Recent efforts have been made to compute city-scale CBFs using final consumption vectors reported in subnational input-output (IO) tables (Stanton et al. 2011). While CBFs incorporate all trans-boundary GHGs relating to local household consumption, local production of exports is allocated out. Such allocation alters the definition of a community, where the geopolitical unit is split in two: local final consumption sectors, and local producers who export goods elsewhere.

Both approaches have their advantages and disadvantages, and neither is complete, in that neither fully accounts for all life cycle supply chains serving both producers and consumers in cities. The TBIF method accounts for life cycle GHGs of essential infrastructures serving cities, but does not account for life cycle GHGs of all other, non-infrastructure goods and services consumed by households or those used in industrial production. The TBIF method recognizes that a city's infrastructure supports both production and consumption activities. Hence it focuses on key trans-boundary infrastructures that provide commuter travel, airline travel, freight, water, energy, food, building materials and waste management services to the entire community as a whole. In contrast, the CBF method ignores the in-boundary and supply chain impacts of commercial–industrial activities that are exported, focusing only on consumption and its supply chains.

The utility of the TBIF has been described in the work of Ramaswami and colleagues (2011). In summary, the TBIF informs community-wide infrastructure planning for sustainability by addressing direct energy use and also embodied energy in infrastructures. The method keeps a community's energy and materials use together (residential and business activity), quantifying community-wide GHG emissions as a whole. The method can link in-boundary energy use and GHG emissions to local air pollution and local health impacts, and is able to track the effects stemming from infrastructure policies across-scale addressing buildings energy supply, transportation, water and wastewater, and waste. By its trans-boundary inclusions, the TBIF method addresses regional cross-sector and cross-scale infrastructure efficiencies, such as mass transit, or expanded tele-presence aimed at reducing air travel. Lastly, supply chain vulnerabilities impacting local economies as a whole are addressed.

ICLEI-USA has gathered a group of technical leaders from business, government, and academia to develop a draft of community-scale GHG emissions accounting and reporting protocol (ICLEI 2012). The protocol recognizes and seeks to address the need for standardized GHG emissions accounting for cities. Four reporting approaches are defined in the protocol framework; basic, expanded community-wide, local government influence, and household consumption. Both the Basic (required) and its associated expanded community-wide report draw upon the TBIF. See figure 1 for a full description of the basic and expanded community-wide reporting frameworks.

Figure 1.

Basic and expanded community-wide reporting frameworks for the International Council for Local Environmental Initiatives United States (ICLEI-USA) community-scale greenhouse gas (GHG) emissions accounting and reporting protocol. LPG = liquefied petroleum gas; WW = wastewater.

The main objective of this article is to evaluate the TBIF method using Delhi, India, as the case study. More specifically, this article (1) describes data availability for implementing the TBIF within Delhi, a rapidly industrializing city; (2) identifies methodological differences between the implementation of the TBIF in Indian versus U.S. cities; and (3) compares broad energy use metrics between Delhi and U.S. cities, as demonstrated by Denver, Colorado, USA, whose TBIF per capita has been shown to be similar to U.S. averages.

Trans-boundary Infrastructure Supply Chain Footprint Method

Description and Data Needs

The TBIF method accounts for in-boundary GHG emissions from buildings (residential, commercial, and industrial), road transportation, industrial processes (i.e., waste emissions, calcination), and the embodied GHG emissions of a city's trans-boundary infrastructure supply chains, including electricity supply, fuel production, water and wastewater treatment, cement production, spatially allocated airline and freight transport, and production of food consumed in the city (see figure 1). The method has been tested in the United States (e.g., Hillman and Ramaswami 2010; Ramaswami et al. 2008), yielding a convergence in per resident GHG emissions from the city to national scale for a set of seven U.S. cities, suggesting that the inclusion of these selected trans-boundary infrastructures generates scale consistency from the city to national levels. Figure 2 illustrates energy and material uses accounted for by the TBIF, along with appropriate benchmarks and associated emission factors (EF). Table 1 illustrates the data needs for benchmarking energy and material use described by the TBIF.

Table 1. Data needs for benchmarking the TBIF method to represent energy and materials use
Activity sectorData needs


  1. VKT = vehicle kilometer traveled; CNG = compressed natural gas; smr = residential square meters; smc = commercial square meters; sm = total square meters; HH = households; GDP = gross domestic product.

Buildings energy use and• Total residential floor area (smr)
 industrial process emissions• Total commercial floor area (million smc)
 • Total floor area per capita (sm/capita)
 • Residential electricity, natural gas, cooking fuels, and heating fuels use
 • Commercial–industrial–government electricity, natural gas, other fuel use
 • Total waste generated in city
Transportation energy use• Allocated daily VKT (VKT/capita/day)
 • Fleet fuel efficiency
 • Volume of gasoline, diesel, and CNG used in road transport
 • Number of enplaned passengers at regional airport (domestic, international)
 • Jet fuel liters loaded into airplanes
 • Percentage of planes fueling at airport
 • Tonnes of long-distance freight and liters of fuel per ton moved
 • Energy used in rail transport
Materials use• Volume of water used (i.e., pumped)
 • Energy used in pumping water
 • Volume of wastewater treated
 • Energy used in wastewater treatment
 • Percentage of water used for residential, commercial, and industrial uses
 • Food consumed/used in the community
 • Cement use in the community
Figure 2.

In-boundary energy and materials use benchmarks, and associated emissions factors (EFs) for the TBIF method. cap = capita; CF = cooking fuel; CO2-eq = carbon dioxide equivalents; EF = emission factor; GDP = gross domestic product; HF = heating fuel; HH = households; kg = kilogram; kWh = kilowatt-hour; LPG = liquefied petroleum gas; m3 = cubic meter; MJ = megajoules; mo = month; t = metric tonnes; PKT = person kilometers traveled; Prod, LCA = production life cycle assessment; smr = residential square meters; smc = commercial square meters; sm = total square meters; km2 = square kilometer; TJ = terajoules; VKT = vehicle kilometers traveled; WW = wastewater; yr = year.

Socioeconomic Profile and Overview of Energy Use and Greenhouse Gases for Delhi, India

India's national population is estimated at 1.155 billion people (World Bank 2010), corresponding to about 17% of the world's population. India's gross domestic product (GDP) is $1,310 billion USD-Real ($3,275 billion USD-PPP),2 roughly 3% of the world's GDP (World Bank 2010), and total primary energy use is estimated at 21 exajoules (EJ) (EIA 2010),3 about 4% of the world's total primary energy use. India's annual growth in primary energy use and GDP are 7% and 8.2%, respectively, relative to trends for the United States of 0.3% and 2.3%, respectively. In Delhi, even greater GDP growth is projected, with annual GDP growth reported at 15.9% (DES 2009).

Delhi is a city-state and the capital of India. Home to almost 18 million people, it boasts a vibrant economy that is poised for continued growth. Spurred by an influx of jobs in information technology (IT), telecommunications, banking, and manufacturing, Delhi has become an attractive place for many, generating a 2009 per resident real GDP that is about twice that of India ($2,415 vs. $1,134 USD-Real) (see table 2). The Delhi government is also initiating a wide range of sustainable infrastructure programs addressing energy use and GHG emissions. Two such examples are the new more-stringent building codes (Energy Conservation Building Code [ECBC] implemented in 2009 [DDE 2009]), and fuel switching of all commercial bus and taxi fleets from gasoline/diesel to natural gas (Indian Supreme Court legislation in 1998 [Mehta 2001]), both of which can help reduce carbon emissions per GDP.

Table 2. Comparisons of key demographic and economic variables in the United States, India, and Delhi
2009United StatesIndiaDelhi


  1. a

    Population statistics sources: U.S. and India (World Bank 2010); Delhi (DCO 2009).

  2. b

    GDP sources: U.S. and India (World Bank 2010); Delhi (DES 2009).

  3. c

    Per capita income sources: U.S. (BEA 2009); India and Delhi (CSO 2009).

  4. d

    Sources for GHG estimates: U.S. (EPA 2011) ; India (MEF 2010); Delhi (estimated in this study).

  5. e

    Primary energy sources: U.S. and India (International Energy Agency); Delhi (estimated).

  6. f

    Data was available for current year only.

  7. *GDP in Real U.S. dollars (USD).

  8. GDP = gross domestic product; USD = U.S. dollars; PPP = purchasing power parity; GHG = greenhouse gas; t CO2-eq = metric tonnes carbon dioxide equivalents; EJ = exajoules.

Population (million)a3071,15517.6
Annual % change0.93%1.4%2.9%
% urbana80.8%29.8%93.2%
% rurala19.2%70.2%6.8%
GDP (billion USD-Real); {billion USD-PPP}b$14,119($1,310); {$3,275}($42.5); {$106}
Annual % change2.3%8.2%15.9%
GDP/capita (USD-Real/capita); {USD-PPP/capita}$45,989($1,134); {$2,835}($2,415); {$6,037}
Annual % change1.4%8.3%12.6%
Income/capita (USD/capita)c$40,947$833$1,965
GHG/capita (t CO2-eq/capita)d21.61.52.4
GHG/GDP (t CO2-eq/million $GDP)*4821,317948
Primary energy (EJ)e104210.53
Annual % change0.3%7%f

Previous research has contributed to some level of energy use and GHG emissions accounting for Delhi. The earliest known GHG emissions research in Delhi was conducted for the baseline year of 1995 by Sharma and colleagues (2002a, 2002b, 2002c, 2002d). That research evaluated GHG emissions from the use of electricity, natural gas, LPG, kerosene, gasoline, and diesel, plus the embodied emissions associated with the production of cement, steel, rice, and milk used in Delhi (Sharma et al. 2002a, 2002b, 2002c, 2002d). A more recent study inventoried Delhi's 2007 GHG emissions from in-boundary activities only (Ghosh 2009).

Data Sources and Results

Indian energy use data are more readily available at the state level than at the city level. Because Delhi is a city-state, considered a union territory among six others in India, energy use data were readily available, which may not be the case for other Indian cities. This section first introduces important demographic trends in Delhi. Then, required data sources and their availability for completing the TBIF in Delhi are discussed, with results presented for demographics and then for the activity sector categories presented above in figure 2 and table 1: (1) building energy use and industrial processes, (2) transportation energy use, and (3) materials use.


Population data for Delhi were obtained from the Directorate of Census Operations (DCO) through the Delhi Statistical Handbook (DSH) (DCO 2009), which reports Delhi's 2009 population equal to 17.6 million people. Household counts in Delhi were last reported by the DCO in 2001. Thus, using home occupancy as reported in 2001 (4.6 people per household), we estimated Delhi's homes to number 3.8 million. Two estimates of population density were obtained: the first was a 2001 (DCO 2009) estimate, equal to 9,340 people per square kilometer (people/km2), and the second was a 2007 estimate equal to 11,463 people/km2 (UN 2010).4

Delhi employment statistics, which were last reported for 2001 (DCO 2009), indicate that the annualized employment growth from 1981 to 1991 and 1991 to 2001 were equal to 5% per annum. Applying the assumption of constant employment growth to the period from 2001 through 2008 yielded an estimated 6.8 million jobs in Delhi in 2009.

Floor areas for residential, commercial, and industrial units in Delhi were not locally available. A literature search yielded national estimates of average urban residential floor areas equal to 46.8 square meters per household (m2/HH) (Thakur 2008),5 and an aggregate India commercial floor area was reported to equal 516 million m2 (Kumar et al. 2010). While assuming that commercial activity occurs in urbanized places, commercial floor areas were apportioned to Delhi by urbanized population, resulting in an estimate for Delhi equal to 25.7 million m2. Industrial floor space is typically difficult to quantify in any community, and was unobtainable in Delhi.

Buildings Energy Use and Industrial Process Emissions

Electricity use in 2009 was reported by the Delhi Electricity Regulatory Commission (DERC 2009). Unlike in the United States, where natural gas is a dominant energy carrier second to electricity, building electricity use in Delhi is followed by a series of other fuels that serve the end use needs of the community, including liquefied petroleum gas (LPG), kerosene, and compressed natural gas (CNG). The use data for these other fuels were obtained from the Ministry of Petroleum and Natural Gas (MPNG 2009) and apportioned to end use sectors using ratios previously estimated for Delhi (Ghosh 2009) (e.g., LPG: 95.9% residential, 3.5% commercial, 0.6% industrial).

Residential Energy Use Benchmarks

Several factors have been shown to contribute to household energy use in India, some of which include home size, home construction material, income, and climate/weather conditions (Pachauri 2004; Pachauri and Jiang 2008). Pachauri (2004) notes that, on average, direct energy use of urban Indian households is two to three times greater than that of rural households.

Electricity use was the dominant end use energy source for Delhi households in 2009, and its monthly use by households is estimated to have been 191 kilowatt hours per household per month (kWh/HH/month) (DERC 2009).6 Nationally, Indian households use 48 kWh/HH/month (IEA 2008). This difference in average household electricity use between Delhi and India is in line with values from Sharma and colleagues (2002c), who estimated Indian urban electricity use to be about three times higher than national averages.

Delhi households typically do not use natural gas or other fuels (e.g., propane) for space heating as is done in the United States, but do use LPG and kerosene for cooking; any coal or biomass use for cooking is not reported in this article. The estimated monthly use of each of the two fuels is 25.3 liters per household (L/HH) for LPG and 3.4 L/HH for kerosene (MPNG 2009).7 This compares to 7.8 L/HH and 4.2 L/HH, respectively, according to national statistics (IEA 2008). Combining these end uses of energy yields an energy end use intensity (EUI) of Delhi residences, estimated at 1,489 megajoules per household per month (MJ/HH/mo).8 India's household EUI is reported at 273 MJ/HH/mo (IEA 2008). These estimates roughly conform to estimates by Pachauri (2004), who notes that urban household energy use is at least triple that of national averages.

Commercial Energy Use Benchmarks

The energy use of commercial buildings consisted of electricity, LPG, CNG, and diesel. Electricity use, excluding use in treating/pumping water and wastewater, was found to equal 5,795 million kWh (DERC 2009), or 225 kilowatt hours per commercial square meter per year (kWh/m2c/yr). Nationally, Gupta and Chandiwala (2011) estimates average commercial electricity use intensity equals 189 kWh/m2c/yr, while estimates provided by the IEA (2010) were 93.6 kWh/m2c/yr for India.

The total end use of the other fuels in commercial buildings is LPG 43 million L, CNG 30.6 million cubic meters (m3), and diesel 15.8 million L (MPNG 2009).9 Combining these energy end uses yields an EUI for Delhi's commercial buildings equal to 923.8 megajoules per commercial square meter per year (MJ/m2c/yr).

Industrial Energy Use Benchmarks

Energy statistics report industries in Delhi used 2,991 million kWh in 2009 (DERC 2009). Other energy end uses by Delhi industries are LPG 6.9 million L, CNG 46.4 million m3, high speed diesel (HSD) 5.9 million L, light diesel oil (LDO) 2.1 million L, and diesel 3 million L (MPNG 2009).

Industrial Process Benchmarks

The Delhi Pollution Control Committee (DPCC) estimates Delhi generates 7,310 tonnes of municipal solid waste (MSW) daily (DPCC 2010),10 amounting to about 0.16 tonnes per resident per year (t/resident/yr), which compares to 0.14 t/resident/yr nationally (Sharholy et al. 2008). About 7% of Delhi's waste is diverted in the form of compost. Additionally there are three ongoing waste-to-energy projects in Delhi that promise to divert close to 15% of today's MSW (DPCC 2010).

Releases of untreated wastewater can also be a source of considerable GHG emissions. Rivers, lakes, lagoons, and the like provide anaerobic conditions for untreated wastewater, resulting in methane (CH4) and nitrous oxide (N2O) production. It is estimated that Delhi captures and treats 63% of its total wastewater produced (MUD 2010). Noting that Delhi treated 1,584 million L of wastewater per day in 2009 (MUD 2010), we estimate the 2009 releases of untreated wastewater total 339,633 million L.

Among the other industrial processes recognized by the Intergovernmental Panel on Climate Change (IPCC) as contributors to GHG emissions, cement production is the most prominent (IPCC 2006a). The Cement Manufacturers Association (CMA) reports no cement production within the boundaries of Delhi, thus providing a basis for incorporating cement as a relevant scope 3 item. No other industrial process emissions were readily identified within Delhi boundaries.

Emissions Factors


Electricity is generated in Delhi at five power plants: three coal powered and two natural gas powered. Their emissions factors (EFs), in kilograms of carbon dioxide equivalents per kilowatt hour (kg CO2-eq/kWh) are 1.16, 1.52, 1.39, 0.59, and 0.36, respectively (Ghosh 2009). Nationally India has two power grids. The first grid is the Integrated Northern, Eastern, Western, and North-Eastern (NEWNE), which has an EF equal to 0.83 kg CO2-eq/kWh. The second is the Southern grid, whose EF is equal to 0.76 kg CO2-eq/kWh. This results in a blended national electricity EF equal to 0.82 kg CO2-eq/kWh (CEA 2009), previously reported to consist of 90% coal, with the remaining 10% made up of natural gas, oil, and wind (MEF 2010). The national electricity EF includes transmission and distribution (T&D) losses (including unauthorized connections), which have been estimated to equal about 24% across India (World Bank 2010). Because the NEWNE regional grid serves Delhi, its electricity EF was used upon the recommendation of ICLEI-South Asia (SA).

Fuel Production and Combustion

The combustion EFs of fuels used in buildings were obtained from the 2007 national India inventory (MEF 2010), and are consistent with those of the IPCC (2006b). The EFs for fuel combustion are natural gas 2.15 kg CO2-eq/m3, LPG 1.6 kg CO2-eq/L, and kerosene 2.7 kg CO2-eq/L. Production EFs of LPG and kerosene were adopted from work by Lewis (1997), since no India-specific data were identified. Those production EFs are reported as LPG 0.26 kg CO2-eq/L and kerosene 0.22 kg CO2-eq/L.

Municipal Solid Waste

The EF from waste landfilling is estimated using IPCC's default methodology (IPCC 2006c):

display math(1)

where MSWT is the total waste generated; MSWF is the fraction sent to landfills; MCF is the CH4 correction factor; DOC is the degradable organic carbon; DOCF is the fraction of DOC dissimilated; F is the fraction of CH4 in landfill gas, with a default value of 0.5; R is the recovery of CH4; and OX is the oxidation factor, with a default value of 0. For the variables requiring specific data relating to Delhi's waste composition, namely MCF, DOC, and DOCF, we turn to the literature. Both Sharma and colleagues (2002b) and Kumar and colleagues (2004) estimate MCF and DOC at 0.4 and 0.15, respectively. Their estimates of DOCF differ, however; where Sharma and colleagues (2002b) report 0.5, Kumar and colleagues (2004) report 0.77. Upon substituting these variables into equation (1), we estimated the range of Delhi's EF from landfilling as 0.4 to 0.6 kg CO2-eq/kg of landfilled waste, and used the average of the two.

Methane and Nitrous Oxide from Released Wastewater

The EF relating to CH4 and N2O production from released untreated wastewater is consistent with the IPCC methodology. The EF for describing CH4 production is

display math(2)

where Cinfluent–COD is the concentration of chemical oxygen demand (COD) in the influent treated wastewater, which for Delhi has been estimated by the DPCC (2010) as an average of all Delhi treatment plants equal to 407 kg COD/million L; B0 is the maximum CH4-producing capacity, and its default value is 0.25 kg CH4/kg COD; MCF is the methane correction factor for rivers and lakes, and its default value is 0.1; and math formula is the methane global warming potential, equal to 24 kg CO2-eq/kg CH4. Multiplying the four terms yields 244 kg CO2-eq/million L of CH4 from Delhi's untreated released wastewater.

The EF for N2O from untreated wastewater releases is adapted from a study by Beaulieu and colleagues (2011), and is written as

display math(3)

where math formula is the concentration of inorganic nitrogen in the influent wastewater. Because Delhi-specific data were not available, we assumed the concentration equaled that of another Indian city, Hyderabad, for which it has been estimated as 52 kilograms of nitrogen per million liters (kg N/million L) (Miller 2011). math formula is the default value of N2O emissions from nitrification and denitrification in rivers—0.005 kilograms of nitrous oxide per kilogram of nitrogen (kg N2O/kg N) (IPCC 2006c). math formula is the nitrous oxide GWP, equal to 298 kg CO2-eq/kg N2O. Multiplying the three terms yields 77 kg CO2-eq/million L of N2O from Delhi's untreated released wastewater.

Transportation Energy Use

Surface Travel Benchmarks

Estimating energy use and GHG emissions from road transport can be challenging in U.S. cities due to the trans-boundary movement of vehicles across multiple cities in a commuter-shed. For example, in the Denver region, consisting of 10 cities (including Denver), 59% of workers commute into Denver and 33% of Denver residents travel outside for work (DRCOG 2007). In this study, because Delhi is a mega city, we can assume the administrative boundaries of Delhi and the commuter-shed overlap, which significantly simplifies the analysis. The assumption was confirmed by finding that only 3% of Delhi's vehicle kilometers travelled (VKT) are trans-boundary (see text below). The latest estimate of Delhi's in-boundary VKT was obtained from the Central Road Research Institute (CRRI), and is reported as 151 million daily VKT for 2009 (CRRI 2009), yielding 8.8 VKT/resident/day. The CRRI study also estimated daily vehicle counts entering and leaving Delhi as 431,246 (inbound) and 464,183 (outbound). To estimate the proportion of VKT associated with trans-boundary traffic, the average Delhi vehicle trip length, estimated to be 10 km (Sahai and Bishop 2010), was applied to either inbound or outbound traffic, therefore estimating that only the equivalent of 3% of Delhi's in-boundary VKT crosses the city boundary. Thus we hypothesize that in megacites, VKTs attributed to trans-boundary traffic may be negligible due to the large amounts of concurrent in-boundary traffic.

The other critical component of the vehicular benchmark is the fuel efficiencies of vehicles in Delhi. Because data on fuel efficiencies are not currently collected by any Indian government agency, estimates from the Automotive Research Association of India (ARAI 2007) were used by Arora and colleagues (2011) in estimating the fuel efficiencies of Indian vehicles (see table 3). We then coupled fuel efficiencies by vehicle type with Delhi's VKT to estimate the fuel used in road transport. Upon allocating the fuel use of outbound vehicle trips, we estimated 2009 fuel use in Delhi road transport: gasoline 1,547 million L, diesel 1,128 million L, and CNG 692 million m3.

Table 3. Surface transport fuel use in Delhi by fuel type and vehicle type
Fuel typeVehicle typeDaily VKT (million)aFuel efficiency* (km/L)bFuel use (million L)Total fuel use, by fuel type


  1. a

    Daily VKT in Delhi retrieved from CRRI (2009).

  2. b

    Average fuel efficiencies within Indian fleet, from Arora and colleagues (2011) and ARAI (2007).

  3. c

    CNG fuel efficiencies shown in liters per cubic meter.

  4. d

    CNG fuel use shown in cubic meters.

  5. *Fuel efficiency is referred to as fuel economy in the United States and reported in equivalent units, miles per gallon.

  6. CNG = compressed natural gas; VKT = vehicle kilometers traveled; km = kilometer; L = liter; m3 = cubic meter.

Gasoline (petrol)Car, small31.113.38251,547 million L
 Car, big13.513.3358 
 Two wheelers54.753.1364 
DieselCar, small8.813.52311,128 million L
 Car, big11.611.9346 
 Light commercial vehicles (LCV)3.35.2231 
 Heavy commercial vehicles (HCV)2.42.8266 
CNGCar, small2.115.4c48.1d692 million m3
 Car, big0.815.4c17.8d 
 Auto (rickshaws)19.630.5c232.3d 

Air Travel Benchmarks

Jet fuel loaded and passenger traffic at Delhi's Indira Gandhi International (IGI) airport was obtained directly from the airport. Jet fuel loaded in 2010 is reported for domestic travel to be 551 million L and for international travel to be 1,214 million L, and enplaned passengers are reported as 8.7 and 4.0 million passengers for domestic and international travel, respectively (DIAL 2010). A passenger survey was conducted at IGI to allocate jet fuel loaded in Delhi based on the proportion of outbound passengers at IGI who were associated with activities in Delhi, either as residents, business travelers, tourists leaving, or visitors to Delhi.

The survey results show that 25% of domestic passengers and 47% of international passengers were traveling through Delhi from another town (see table 4). Thus 76% of domestic passengers and 53% of international passengers can be deduced to have Delhi-related travel, which was used to allocate jet fuel loaded to Delhi. Allocating jet fuel and passengers to Delhi yields 414 and 644 million L for domestic and international travel, respectively, thus resulting in 56 and 275 L/enplaned passenger for domestic and international travel, respectively. Of the total jet fuel loaded at IGI, only the domestic portion was incorporated into Delhi's TBIF as required by international protocols (UNFCCC 2006).

Table 4. Summary of relevant results from the airport survey conducted at the Delhi International Airport (total surveyed = 111 travelers: 52 domestic travelers and 59 international travelers)
  % Responses
QuestionAnswer choiceDomestic terminal (n = 52)International terminal (n = 59)
1. Are you a resident ofa. Yes27%27%
 Delhi?b. No73%73%
2. If not a resident ofa. Business or work-related trip in Delhi35%5%
 Delhi, are you leaving after a…b. Holiday or other special occasion in Delhi6%0%
 c. Visited friends or relatives in Delhi6%15%
 d. Sightseeing tour/vacation in Delhi2%5%
 e. None of the above: I am just passing through Delhi from another city or town25%47%

Rail Travel Benchmarks

We used India's national GHG emissions inventory to determine that emissions from railways, mostly diesel combustion, constitute 0.4% of the country's GHG emissions (MEF 2010). A lack of data and the relatively lower importance in terms of total national GHG emissions guided us to ignore GHG emissions from railways in Delhi at this stage. With a new local commuter rail being installed in Delhi, future work may incorporate GHGs from rail by combining the energy use of Indian railways (IRFCA 2006b), rail passenger kilometers traveled (PKT), and goods transported by rail (World Bank 2010).

Emissions Factors

The combustion EFs of fuels used in transportation within Delhi are consistent with those from the IPCC (2006) and equal to those used in India's national GHG inventory (MEF 2010). The EFs from fuel combustion are gasoline 2.4 kg CO2-eq/L, diesel 2.9 kg CO2-eq/L, and jet fuel 2.7 kg CO2-eq/L. The EF associated with diesel production was retrieved from the work of Whitaker (2007), who estimated an EF from diesel production in India equal to 0.5 kg CO2-eq/L. Because the distillation temperature of diesel occurs within a similar range for that of jet fuel kerosene, 200°C–300°C, it was assumed that both diesel and jet fuel have a similar production EF, as has been previously assumed (Hillman and Ramaswami 2010; Kennedy et al. 2009). Production EFs for gasoline and CNG in India were not available, so the following assumptions were made. For gasoline, because distillation occurs at lower temperatures than for diesel, the worst-case EF was assumed to be equal to diesel; and CNG was assumed to equal the median value of those reported in the work of Kennedy and colleagues (2009) (see table 5).

Table 5. Illustrating the emissions factors (EFs) of transportation fuels in Delhi. We calculate the combustion-to-life cycle EF ratio of each fuel and compare it to the range of values estimated by Kennedy and colleagues (2009). Note: The EF applied in Delhi is in line with ranges estimated by others
FuelDirect PTW GHG (kg CO2-eq/L)Processing WTP GHG (kg CO2-eq/L)Lifecycle WTW GHG (kg CO2-eq/L)PTW/WTW (Delhi)PTW/WTW (Kennedy et al. 2009)


  1. *CNG reported in kg CO2-eq/m3.

  2. PTW = pump-to-wheels (combustion); WTP = wells-to-pump (production); WTW = wells-to-wheels (life cycle), which equals WTP + PTW; CNG = compressed natural gas; kg CO2-eq = kilograms of carbon dioxide equivalents.

Jet fuel2.70.53.284%Assumed diesel

Embodied Energy of Materials Use

Embodied energy incorporated in the TBIF includes that for wastewater treatment and pumping, water treatment and pumping, food production, and cement production, since these activities are not already counted in in-boundary GHG emissions described in previous sections. Although wastewater and water treatment occurs within Delhi, subtracting these energy uses from the above estimates allows us to clearly illustrate embodied energy used in wastewater and water operations.

Embodied Energy of Materials Benchmarks

Wastewater treatment in Delhi is tracked and reported by the Delhi Jal Board, which treated 1,584 million L/day of wastewater in 2009 (MUD 2010), using a total of 40 million kWh of electricity annually (treatment [T] = 17 million kWh, pumping [P] = 23 million kWh) (DJB 2011 2011). Municipal treated water supply totaled 3,125 million L/day (MUD 2010), using a total of 266 million kWh of electricity annually (T = 242 million kWh, P = 24 million kWh) (DJB 2011 2011).

For estimating average food consumption by Indian households, Miller and Ramaswami (2012) used statistics from the United Nations Food and Agriculture Organization (FAO), thus resulting in 3,616 kg of food/HH (or 0.78 t of food/resident), thereby estimating the 2009 food supply to Delhi as 13.8 million t. This likely underestimates all food used in Delhi, as it excludes food in commercial and tourist establishments, which may be a large portion of the city's economy. Further, there are an estimated 45,285 nonmilk heads of cattle, 45,760 milk heads of cattle, and 304,655 heads of buffalo within Delhi boundaries (DAH 2010), which we use in the next subsection for estimating direct CH4 emissions from enteric fermentation within Delhi boundaries. These in-boundary estimates associated with milk-producing cattle are subtracted from the trans-boundary GHG emissions from food production.

As previously discussed, there is no cement production within Delhi boundaries (CMA 2010), thus confirming that Delhi cement flows can be treated as trans-boundary. Community-wide cement use in Delhi was obtained from the Cement Manufacturers Association (CMA) and is estimated at 0.24 t/capita/yr (CMA 2010).

Emissions Factors

Water and wastewater are supplied from within Delhi, and thus relevant energy use has been subtracted from Delhi's in-boundary energy use total, avoiding double counting. End use energy intensity used in 2009 for treating and pumping wastewater and water were obtained from work by DJB (2011 2011) and MUD (2010). The resulting ratios of energy to water are 0.03 watt-hours per liter (Wh/L) of treated wastewater, 0.04 Wh/L of pumped wastewater, 0.21 Wh/L of treated water, and 0.02 Wh/L of pumped water.

The food EF was retrieved from the work of Miller and Ramaswami (2012), who estimated a per unit weight EF from Indian food production (agriculture only). Their food EF quantifies direct CH4 and N2O emissions from Indian agriculture, eliminating double counting of the energy used in processing or transporting food. The food EF, including emissions from cattle, is 0.45 tonnes of carbon dioxide equivalent per tonne of food (t CO2-eq/t food).

This study also considered direct CH4 emissions from enteric fermentation in Delhi. GHG emissions per cattle in Delhi were retrieved from the work of Sharma and colleagues (2002b), who estimated the EF from non-milk-producing cattle as 525 kilograms of carbon dioxide equivalents per head per year (kg CO2-eq/head/yr), for milk producing cattle as 966 kg CO2-eq/head/yr, and for buffaloes as 1,155 kg CO2-eq/head/yr, yielding GHG emissions from cattle within Delhi boundaries of 419,855 t CO2-eq/yr. Lastly, to avoid double counting we subtract cattle GHG emissions from the above food EF, resulting in a new food EF (less in-boundary cattle) equal to 0.42 t CO2-eq/t food.

The cement EF is well documented and obtained from the literature. In this study we applied an Indian EF from cement production equal to 0.93 t CO2-eq/t cement (Hendriks et al. 2004).


In this article we presented the methodology and results from applying the TBIF GHG emissions accounting method in the rapidly developing city of Delhi, India. The objectives were to (1) describe data availability for implementing the TBIF in Delhi, (2) identify methodological differences between India and U.S.-based implementation of the TBIF, and (3) compare broad energy use metrics between Delhi and U.S. cities, as represented by Denver, which has previously been shown to be similar to U.S. averages.

Multiplying Delhi's 2009 material and energy flows with associated EFs resulted in total TBIF GHG emissions equal to 40.3 million t CO2-eq. Normalizing by population, Delhi's TBIF GHG emissions are 2.3 t CO2-eq/capita; as expected, they are higher than the 1.5 t CO2-eq/capita reported nationally (MEF 2010), since Delhi represents 1.5% of India's population. Of Delhi's 2009 TBIF GHG emissions, in-boundary activities represented 68% (or 27.3 million t CO2-eq) and trans-boundary activities 32% (or 13 million t CO2-eq). The buildings sector (including residential, commercial, and industrial) represented 42% of Delhi's GHG emissions. GHGs from road transportation represented 21%, waste 2.7%, water and wastewater pumping and treatment 0.6%, and cattle 1% (see figure 3).

Figure 3.

Delhi's expanded greenhouse gas (GHG) emissions footprint in 2009. In-boundary GHG emissions are represented by solid wedges and trans-boundary GHG emissions are represented by hatched ones. t CO2-eq = metric tonnes of carbon dioxide equivalents; mill $ GDP = million dollars of gross domestic product; Bldgs = buildings; WW = wastewater.

The TBIF method was found to be very useful for measuring a comprehensive GHG footprint for Delhi. Most of the required data for applying the TBIF for Delhi were found to be available, and it is possible that this could have been a result of high levels of government reporting, as Delhi is a city-state. There were some methodological differences, such as the airport survey, that were a result of data constraints, although the method was mostly replicated. In fact, the method applied for Delhi helped to identify clear data needs and knowledge gaps where supplementary primary data collection is needed. For example, in the United States, the method has used regional transportation models for allocating jet fuel use to multiple cities served by a single regional airport. In Delhi, the absence of a transportation model required the use of airport surveys to allocate the total jet fuel used at the Delhi International Airport to Delhi. In many ways, implementing the TBIF was easier in a large megacity such as Delhi. Trans-boundary VKT was found to be a small contribution (3%) of in-boundary VKT, thus origin-destination allocation of travel between cities in a commuter-shed may not be needed for mega-cities. Further, the CMA reports annual cement use in city-states such as Delhi, while obtaining these data has been challenging in U.S. cities, as will likely be the case in other Indian cities that are not city-states.

Comparing broad metrics across two distinct cities—Delhi and Denver—presents compelling results. Delhi's per capita GHG emissions are higher than India's (2.4 vs. 1.5 t CO2-eq/cap), reflecting low urbanization levels in India. Both Ramaswami and colleagues (2008) and Hillman and Ramaswami (2010) note that Denver's GHG emissions are fairly close to the U.S. national average, at about 25 t CO2-eq/capita, due to 80% U.S. urbanization, that is, 80% of the population living in urban areas. Delhi's per capita GHG emissions are almost a factor of ten lower than Denver's, which can be explained by a multitude of factors. For example, Delhi's residential primary energy use of 3,693 MJ/HH/mo is a factor of three lower than Denver's (10,551 MJ/HH/mo), and road transportation travel in Delhi (8.8 VKT/cap/day) is about four times less than Denver's (39 VKT/cap/day). Most notable are the differences in commercial–industrial energy end use, which is significantly lower in Delhi (2,064 MJ/capita/yr) compared to Denver (76,166 MJ/capita/yr). Similarly, commercial floor area per capita is much less in Delhi than Denver (1.46 m2/capita vs. 36.7 m2/capita). Even though the data suggest much less commercial activity in Delhi versus Denver, the economic GHG intensity provides additional insights.

Delhi's economic GHG intensity is twice as large as Denver's, at 948 tonnes of carbon dioxide equivalents per million dollars of gross domestic product (t CO2-eq/million $ GDP) versus 413 t CO2-eq/million $ GDP, respectively. Such a difference may be attributed to economic structure, where Denver is predominantly a tertiary sector producer and Delhi is a secondary and tertiary sector producer. Other notable differences are shown in table 6. Delhi's GDP/capita is about ten times less than Denver's ($6,037 USD/capita vs. $57,560 USD/capita). In terms of population, Delhi is significantly denser than Denver (9,340 persons/km2 vs. 1,463 persons/km2). Also, homes in Delhi are smaller than homes in Denver (46.8 m2/HH vs. 102.8 m2/HH).

Table 6. Comparing energy and material use, and demographic metrics for Delhi, India, and Denver, Colorado, USA. Also shown are the electricity emissions factor (EF) for both cities and countries
Activity sectorMetricDelhi, India (2.3 t CO2-eq/cap)Denver, CO, USA j (25 t CO2-eq/cap)


  1. a

    DERC (2009).

  2. b

    MPNG (2009).

  3. c

    DPCC (2010).

  4. d

    CRRI (2009).

  5. e

    DIAL (2010).

  6. f

    MUD (2010).

  7. g

    CMA (2010).

  8. h

    DES (2009).

  9. i

    DCO (2009).

  10. j

    Hillman and Ramaswami (2010).

  11. k

    BEA (2009).

  12. #Calculated.

  13. *Estimated, and may not represent the most accurate statistic.

  14. Electricity EF: No brackets represent local EF; {brackets} represent national EF; n/a = not applicable.

  15. cap = capita; CO2-eq = carbon dioxide equivalents; GDP = gross domestic product; HH = households; kg = kilogram; kWh = kilowatt-hour; LPG = liquefied petroleum gas; m3 = cubic meter; MJ = megajoules; mo = month; t = metric tonnes; smr = residential square meters; smc = commercial square meters; sm = total square meters; km2 = square kilometer; VKT = vehicle kilometers traveled; WW = wastewater; yr = year.

Buildings energy useResidential intensity:
 and industrial kWh/HH/mo191a545
 process m3/HH/mon/a124
  L LPG/HH/mo25.3bn/a
  L kerosene/HH/mo3.4bn/a
  Total MJ/HH/mo (end use)1,489#6,728
  Total primary (MJ/HH/mo)3,693#10,551#
 Commercial–industrial intensity:
  Other stationary fuels (MJ/GDP/yr)0.13#0.78#
  Total MJ/GDP/yr (end use)0.87#1.32#
  Total MJ/capita/yr (end use)2,064#76,166#
  Total primary (MJ/GDP/yr)3.5#2.4#
 Industrial process (t waste/capita/yr)0.16c1.1
 Electricity EF (kg CO2-eq/kWh)0.82 {0.83}0.75 {0.64}
Transportation energySurface travel intensity (VKT/capita/day)8.8d38.6
 useAir Travel: liters-jet fuel/enplaned passenger (domestic)56e72
Materials use, andWater: treated water/WW (1000 liters/capita/yr)95f560
 demographicsCement: t cement/capita/yr0.24g0.50
  GDP/capita ($/capita)$6,037h$57,560k
  Total local population (capita)17,601,000i579,744
  Population density (capita/km2)9,340i1,463
  Total homes (HH)3,815,104i256,524
  Residential floor area (smr/HH)46.8*102.8
  Total commercial floor area (million smc)25.7*21.3
  Total floor area per capita (sm/cap)10.1#74.5
  Total city area (km2)1,886#396

As shown, the TBIF method can have important environmental and policy implications for Delhi and other rapidly industrializing cities. The TBIF method shows an additional 32% of Delhi's GHG emissions attributed to trans-boundary activities, thereby suggesting innovative cross-sector strategies toward urban sustainability, particularly in electricity generation and the building materials and cement sectors. Comparing Delhi to Denver, supply chain GHGs from cement use in construction contributed 10% to Delhi's TBIF, versus only 2% in Denver (Ramaswami et al. 2008); in contrast, waste and wastewater GHGs were a lower proportion in Denver (at 1%) versus Delhi (at 3.3%). These data suggest that other construction materials not studied here may also be a significant part of Delhi's TBIF. The TBIF for Delhi shows that both waste management and material exchange symbiosis can be important in reducing the TBIF of cities in rapidly industrializing countries.


Abel Chavez was supported by the U.S. National Science Foundation's Integrative Graduate Education and Research Traineeship (IGERT; grant no. DGE-0654378). Warm appreciation is given to the International Council for Local Environmental Initiatives–South Asia (ICLEI-SA) staff for all of their assistance. We also thank Dr. Muthukrishnan from the Delhi airport, who was instrumental in our data collection efforts and provided personnel to carry out the airport survey. Leslie Miller was also extremely resourceful throughout the data collection efforts, and her assistance was critical.


  1. 1

    The organization responsible for this framework was originally called the International Council for Local Environmental Initiatives, but it changed its name in 2003 to ICLEI – Local Governments for Sustainability.

  2. 2

    “Real” denotes constant dollars that are adjusted for inflation and price differences. PPP denotes purchasing power parity.

  3. 3

    One exajoule (EJ) = 1018 joules (J, SI) ≈ 9.48 × 1014 British Thermal Units (BTU).

  4. 4

    One square kilometer (km2) = 100 hectares (ha) ≈ 0.386 square miles ≈ 247 acres, so 9,340 people per square kilometer corresponds to about 24,197 people per square mile.

  5. 5

    One square meter (m2) is equivalent to 10.76 square feet (ft2), so 46.8 m2 per household corresponds to 503.8 ft2.

  6. 6

    One kilowatt-hour (kWh) ≈ 3.6 × 106 joules (J, SI) ≈ 3.412 × 103 British Thermal Units (BTU).

  7. 7

    One liter (L) = 0.001 cubic meters (m3, SI) ≈ 0.264 gallons (gal).

  8. 8

    One megajoule (MJ) = 106 joules (J, SI) ≈ 239 kilocalories (kcal) ≈ 948 British Thermal Units (BTU).

  9. 9

    One cubic meter (m3, SI) = 103 liters (L) ≈ 264.2 gallons (gal).

  10. 10

    The term tonne refers to metric ton. One tonne (t) = 103 kilograms (kg, SI) ≈ 1.1 short tons.


  • Abel Chavez is a postdoctoral fellow at the University of Colorado Denver in Denver, Colorado, USA. At the time of submission of this article,

  • Anu Ramaswami served as professorof environmental engineering and director of the Center for Sustainable Infrastructure Systems (CSIS) at the University of Colorado Denver. She is presently the Denny Chair Professor of Science, Technology and Public Policy at the Humphrey School of Public Affairs, University of Minnesota, in Minneapolis, Minnesota, USA.

  • Dwarakanath Nath is senior scientific officer at the Department of Environment–Delhi in Delhi, India.

  • Ravi Guru is energy and climate senior project officer at the International Council for Local Environmental Initiatives South Asia (ICLEI-SA) in Delhi, India.

  • Emani Kumar is executive director at ICLEI-SA in Delhi, India.