India is currently experiencing degraded air quality, and
future economic development will lead to challenges for air quality management.
Scenarios of sectoral emissions of fine particulate matter and its precursors
were developed and evaluated for 2015–2050, under specific pathways of
diffusion of cleaner and more energy-efficient technologies. The impacts of
individual source sectors on PM
India hosts the world's second largest population (UNDP, 2017), but accounts
for only 6 % of the world's total primary energy use (IEA, 2015). However,
India is an emerging economy with significant growth in a multitude of
energy-use activities in industry and transport sectors, as well as in
residential, agricultural and informal industry sectors (Sadavarte and
Venkataraman, 2014; Pandey et al., 2014). With expansion in power generation
(CEA, 2016) and industrial production (Planning Commission, Government of
India, 2013), emissions from these sectors were estimated to have increased
about 2-fold between 1995 and 2015 (Sadavarte and Venkataraman, 2014). There
is a steady demand for motorized vehicles for both personal and public
transport, with an increase in ownership of motorized two-wheeler
motorcycles and scooters and four-wheeler cars (MoRTH, 2012), in both rural
and urban areas. Traditional technologies, and the use of solid biomass
fuels, are widespread in the residential sector (cooking with biomass fuel
cook stoves and lighting with kerosene wick lamps), the agricultural sector
(open burning of agricultural residue for field clearing) and the informal
industry sector (brick production, processing of food and agricultural
products). Ambient PM
Air quality is a public health issue of concern in India. According to the
World Health Organization (WHO), 37 cities from India feature in a global
list of 100 world cities with the highest PM
Strategies for mitigation of air pollution require understanding of pollutant
emissions, differentiated by emitting sectors and by sub-national regions,
representing both present-day conditions and future evolution under different
pathways of growth and technology change. Future projections of emissions,
for climate relevant species, are available in the representative
concentration pathway (RCP) scenarios (Fujino et al., 2006; Clarke et al.,
2007; van Vuuren et al., 2007; Riahi et al., 2007; Hijioka et al., 2008),
more recently for the Shared Socioeconomic Pathway (SSP) scenarios (Riahi
et al., 2017; Rao et al., 2017), while primary PM
Section 2 discusses the development of the emission inventory, disaggregated
by sector, for the year 2015 and future projections to 2050; Sect. 3
describes the GEOS-Chem model, the simulation parameters and evaluation;
Sect. 4 discusses simulated PM
An emission inventory was developed for India for the year 2015, based on
an engineering model approach using technology-linked energy-emissions
modelling adapted from previous work (Pandey and Venkataraman, 2014; Pandey
et al., 2014; Sadavarte and Venkataraman, 2014), to estimate multi-pollutant
emissions including those of
Description of source categories and sensitivity simulations.
The inventory disaggregates emissions from technologies and activities in all major sectors. Plant-level data (installed capacity, plant load factor and annual production) are used for 830 individual large point sources, in heavy industry and power generation sectors, while light industry activity statistics (energy consumption, industrial products, solvent use, etc.) are from a sub-state (or district) level (CEA, 2010; CMA, 2007a, b, 2012; MoC, 2007; FAI, 2010; CMIE, 2010; MoPNG, 2012; MoWR, 2007). Technology-linked emission factors and current levels of deployment of air pollution control technologies are used. Vehicular emissions include consideration of vehicle technologies, vehicle age distributions and super-emitters among on-road vehicles (Pandey and Venkataraman, 2014). Residential sector activities comprise cooking and water heating, largely with traditional biomass stoves, lighting, using kerosene lamps and warming of homes and humans, using biomass fuels. Seasonality in emissions from the residential sector is considered for water heating and home warming. The “informal industries” sector includes brick production (in traditional kiln technologies like Bull's trench kilns and clamp kilns, using both coal and biomass fuels) and food and agricultural product processing operations (like drying and cooking operations related to sugarcane juice, milk, food grain, jute, silk, tea and coffee). In addition, monthly mean data on agricultural residue burning in fields, a spatio-temporally discontinuous source of significant emissions, were calculated using a bottom-up methodology (Pandey et al., 2014). Spatial proxies used to estimate gridded emissions over India are described in Table S1 in the Supplement.
India emissions for 2015 of PM
Description of future scenarios.
National emissions of particulate matter and precursor gases for
2015 (Mt yr
Detailed tabulations of 2015 emissions of each pollutant at the state level are provided in Table S2 in the Supplement. Uncertainties in the activity rates, calculated analytically using methods described more fully in previous publications (Pandey and Venkataraman, 2014; Pandey et al., 2014; Sadavarte and Venkataraman, 2014), are shown in Table S3 in the Supplement.
We develop and evaluate three future scenarios which extend from 2015 to 2050, which are likely to bound the possible amplitude of future emissions, based on the expected future evolution of sectoral demand, following typical methods in previous studies (Cofala et al., 2007; Ohara et al., 2007). These include a reference (REF) scenario and two scenarios (S2 and S3) representing different levels of deployment of high-efficiency, low-emissions technologies (Table 2). The scenarios capture varying levels of emission control, with no change in current (2015) regulations, corresponding to very slow uptake of new technology (REF), adoption of promulgated regulations, corresponding to effective achievement of targets (S2) and adoption of ambitious prospective regulations, corresponding to those well beyond promulgated regulations (S3). In both S2 and S3, despite expanding sectoral demand, there is reduced energy consumption from adoption of clean energy technologies, at different levels.
The methodology for emission projection includes estimation of future evolution in (i) sectoral demand, (ii) technology mix, (iii) energy consumption and (iv) technology-linked emission factors (Fig. S1 in the Supplement). Activity levels in future years by source category (e.g. GWh installed capacity in power, vehicle-kilometres travelled in transport, industrial production, e.g. in tons, population of users in residential areas) were apportioned to various technology divisions, using an assumed evolving technology mix, for three different scenarios. Activity at the technology division level was used to derive corresponding future energy (and fuel) consumption and related emissions using technology-based emission factors.
With 2015 as the base year, growth rates in sectoral demand were identified for thermal power plants, industries, residential, brick kilns and informal industries, on-road transportation and agricultural sectors for 2015–2030 and 2030–2050 (Table S4 in the Supplement). Sectoral growth levels, estimated as ratios of 2050 to 2015 demand, were 5.1, 3.8, 3.2, 1.3 and 1.4 respectively, for the building sector, electricity generation, heavy industries, residential sector and agricultural residue burning, the largest growth being in the building and electricity generation sectors (Fig. S2 in the Supplement).
Table 2 shows regulation levels for different sectors under the three
scenarios, through to 2050. The REF and S2 scenarios capture both energy
efficiency and emissions control, continuing under current regulation, or
broadly under promulgated future policies. The S2 scenario assumes shifts to
non-fossil generation which would occur under India's Nationally Determined
Contribution (India's NDC, 2015) in the power sector, consistent with a shift
to 40 % renewables including solar, wind and hydropower by 2030 (India's
NDC, 2015). The NDC goals of India are suggested to be realistic (CAT, 2017;
Ross and Gerholdt, 2017), with achievement of non-fossil share of power
generation projected to lie between 38 and 48 % by 2030, as well as
adoption of tighter emission standards for desulfurization and
de-
However, in the S3 scenario, adoption of ambitious regulations, well beyond those currently promulgated, is assumed. This includes very significant shifts to non-fossil-power generation (Anandarajah and Gambhir, 2014; Shukla and Chaturvedi, 2012; Level 4, IESS, Niti Aayog, 2015), a near-complete shift to high-efficiency industrial technologies (MoP, 2012, Level 4, IESS, Niti Aayog, 2015), a large public vehicle share (NITI Aayog, 2015), energy efficiency improvements in engine technology (MoP, 2015), a large share of electric and compressed natural gas (CNG) vehicles (NITI Aayog, 2015), a complete switch to LPG/PNG or biogas or high-efficiency gasifier stoves for residential cooking and heating (Level 4, IESS, Niti Aayog, 2015) and to solar and electric lighting (National Solar Mission, 2010) by 2030 and a significant (by 2030) and complete (by 2050) phase-out of agricultural residue burning, through a switch to mulching practices (Gupta, 2014). Further details of the shift in technologies can be found in Table S5 and related discussion in the Supplement (see Sect. S2.3).
As alluded to earlier, there is a reduction in total energy consumption in
future years, despite increase in activity, in scenarios S2 and S3, which
assumes the large deployment of high-efficiency energy technologies. The
projected energy demand under the three scenarios (Fig. S3, Sect. S2.4) is in general agreement with published work (Anandarajah
and Gambhir, 2014; Chaturvedi and Shukla, 2014; Parikh, 2012; Shukla et al.,
2009), 95 to 110 EJ for reference scenarios (Parikh, 2012; Shukla
and Chaturvedi, 2012) and 45–55 EJ for low carbon pathways (Anandarajah and
Gambhir,
2014; Chaturvedi and Shukla, 2014) in 2050. Projections of
Technology-based emission factors, for over 75 technology/activity divisions,
are described in previous publications (Pandey et al., 2014; Sadavarte and
Venkataraman, 2014). In addition to fuel combustion, emissions are estimated
from industrial “process” activities predominant in industries such as
those producing cement and non-ferrous metals, and refineries producing iron
and steel (Table S8, Sect. S2.5). In fired-brick production, recently
measured emission factors for this sector of PM
Sectoral emission of fine
The net effect of scenario-based assumptions is that under the REF scenario, emissions are projected to increase steadily over time. Under the S2 scenario, they are also projected to increase but at a slower rate. Only under the most ambitious scenario, S3, are appreciable reductions in emissions of the various air pollutants expected.
Sectoral emission of fine
Emissions of PM
Emissions of
Anthropogenic dust (Philip et al., 2017), defined here as mineral
constituents of pollution particles, including coal fly ash and mineral
matter in waste burning and biomass burning emissions, contributes about
30 % of Indian PM
Emission datasets for India in global emission inventories have been
developed either through a combination of regional inventories for specific
base years (Janssens-Maenhout et al., 2015) or using integrated assessment
models, e.g. the GAINS model (Amann et al., 2011), to generate scenarios of
air pollutants (Klimont et al., 2009, 2017, 2018; Purohit et al., 2010; Stohl
et al., 2015). Indian emissions for 2008 and 2010 under the HTAP_v2
framework (Janssens-Maenhout et al., 2015) originate from the MIX inventory
(M. Li et al., 2017), based on earlier Asia
inventories like INTEX-B (Lu et al., 2011; Lu and Streets, 2012) and REAS
(Kurokawa et al., 2013). Inconsistencies are reported from merging datasets,
calculating different pollutants using differing assumptions (M. Li et al.,
2017). The datasets do not include some important regional emission sources
like the open burning of agricultural residue (Janssens-Maenhout et al.,
2015). Recent global emissions from ECLIPSE V5 (Stohl et al., 2015;
Future emissions of particulate matter (PM
For
For particulate matter species, the GAINS model estimates lower 2015
emissions, mostly because of the differences for residential use of biomass
as well as emissions from open burning. However, considering the
uncertainties associated with the quantification of biomass use and emission
factors (e.g. Bond et al., 2004; Klimont et al., 2009, 2017; Venkataraman et
al., 2010) the differences are acceptable. The future evolution of emissions
of BC and OC shows similar features among the studies with S2 comparable to
ECLIPSE V5a-CLE and S3 to GAINS-WEO2016-NPS; however the S3 scenario shows a
much stronger reduction due to faster phase-out of kerosene for lighting and
stronger reduction of biomass used for cooking. The latter feature is
especially visible for emissions of OC (Fig. 2d, g). For total PM
Emissions of NMVOCs (Fig. 3g) monotonically increase in ECLIPSE V5a-CLE, becoming higher than those in S2 by 2030, which, however, mimic those in GAINS-WEO2016-NPS through to 2050. While there is also a fairly large difference in estimate for the base year (mostly due to residential combustion of biomass, open burning and solvent use sector), obviously the assumptions about the future policies are different as both the ECLIPSE V5a and IEA studies include more conservative assumptions about the reduction of biomass use and eradication of open burning practices, while at the same time they included continued growth in industrial emissions, i.e. solvent applications. Further analysis of differences between the S2 scenario and the ECLIPSE V5a-CLE and GAINS-WEO2016-NPS is shown in the Supplement (Fig. S5).
Further, the emission projections were also compared with emissions estimated
in the four representative concentration pathway (RCP) scenarios adopted by
the IPCC as a common basis for modelling future climate change (Fujino et
al., 2006; Clarke et al., 2007; van Vuuren et al., 2007; Riahi et al., 2007;
Hijioka et al., 2008). The RCP scenarios were designed to represent a range
of possible future climate outcomes in terms of radiative forcing watts per
square metre (W m
The emissions were input in the GEOS-Chem model (
In addition to the emissions described in Sect. 2.2.2, other emissions such
as open burning, except agricultural residue burning which includes forest
fires, were derived from the global GFED-4s database (Akagi et al., 2011;
Andreae and Merlet, 2001; Giglio et al., 2013; Randerson et al., 2012; van
der Werf et al., 2010). In addition to the species in this inventory, ammonia
or
The South Asia nested version of GEOS-Chem used here was developed by
Sreelekha Chaliyakunnel and Dylan Millet (both of the University of
Minnesota) to cover the area from 55 to 105
Model evaluation by
Evaluation of model performance (NMB) in capturing seasonal variation in chemical species concentrations at two sites in India.
To estimate the impacts of individual sources, simulations were made using
total emissions from all sources, along with sensitivity simulations (Table 1) for major sources. Sources included in the standard simulation, however,
that were not separately addressed in sensitivity simulations, termed
“other”, include residential lighting with traditional kerosene lamps and informal
industry (food and agro-product processing). Primary particulate matter is
largely composed of carbonaceous constituents (black carbon and organic
matter) and mineral matter. Mineral matter from combustion and industry,
calculated as the difference between emitted PM
The GEOS-Chem simulations made here include those for primary aerosol
emissions (secondary sulfate, nitrate and ammonium) and secondary organic
aerosol, going beyond previous simulations made on regional scales over
India (e.g. Sadavarte et al., 2016), which were limited to secondary
sulfate and a smaller list of sources in the emissions inventory,
addressing only a few months in the year. Model-predicted concentrations of
PM
Evaluation was also explored against monthly mean chemical composition
measurements (Fig. 5) at a regional background site (Bhopal, 23.2
Simulated PM
As discussed earlier, NMVOC emissions from India were taken from a recent
technology-linked inventory, deployed in WRF-CAMx and evaluated with
satellite and in situ observations (Sarkar et al., 2016). However,
uncertainties still remain to be addressed in the calculation of secondary
PM
We find that ambient PM
Simulations with the REF scenario emissions (Fig. 6b, c), show significant
increases in annual mean PM
We further examine which increases or decreases in PM
Population-weighted mean ambient PM
A similar picture was seen in 2050 as well, with very significant increases
under the REF scenario in all states, leading to extreme PM
The simulated change in sectoral contribution to population-weighted
PM
Percentage contribution to ambient PM
Percentage contribution of
In 2015, among source sectors, the single largest contributor to ambient
PM
In 2050, future source contributions are dominated by power plant coal and
industrial coal, in both REF and S2 scenarios, followed by residential
biomass. In both REF and S2 scenarios (Figs. 2 and 3) expansion in
electricity generation and industry overtakes emissions offsets, leading to
1.5–2 and 1.75–3 times emission increases, respectively, in emissions of
PM
Interestingly, the influence of residential biomass emissions on PM
The PM
Changes in source contributions to PM
Overall, sources significantly influencing PM
This work represents the most comprehensive examination to date of a
systematic analysis of source influence, including all sources, on present
and future air pollution on a regional scale over India. Elevated annual mean
PM
If no action is taken, population exposure to PM
Gridded data files of emissions and PM
The authors declare that they have no conflict of interest.
This article is part of the special issue “Global and regional assessment of intercontinental transport of air pollution: results from HTAP, AQMEII and MICS”. It is not associated with a conference.
Partial support for this work was provided by the Health Effects Institute, Boston. Kushal Tibrewal acknowledges a PhD assistantship from the NCAP-COALESE grant of the Ministry of Environment Forests and Climate Change, India. Dylan B. Millet and Sreelekha Chaliyakunnel acknowledge support from NASA (no. NNX14AP89G), the NSF (no. AGS-1148951) and the Minnesota Supercomputing Institute. We acknowledge Sarath Guttikunda, Co-director, Urbanemissions.info, Goa, India, for the datasets of present-day and future emissions from waste burning and urban fugitive dust. Edited by: Frank Dentener Reviewed by: Sachin D Ghude and one anonymous referee