Simulating ozone concentrations using precursor emission inventories in Delhi – National Capital Region of India

Accepted Manuscript Simulating ozone concentrations using precursor emission inventories in Delhi – National Capital Region of India Sumit Sharma, Mukesh Khare PII:

S1352-2310(16)30968-2

DOI:

10.1016/j.atmosenv.2016.12.009

Reference:

AEA 15071

To appear in:

Atmospheric Environment

Received Date: 19 May 2016 Revised Date:

1 December 2016

Accepted Date: 2 December 2016

Please cite this article as: Sharma, S., Khare, M., Simulating ozone concentrations using precursor emission inventories in Delhi – National Capital Region of India, Atmospheric Environment (2017), doi: 10.1016/j.atmosenv.2016.12.009. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

3

Sumit Sharma1,2, Mukesh Khare3*

4

1 Research Scholar, Civil Engineering Department, Indian Institute of Technology, Hauz Khas, New Delhi-110 016, India

5

2 Fellow, Centre for Environmental Studies, The Energy and Resources Institute, IHC Complex, Lodhi Road New Delhi – 110 003, India, email- [email protected]

6

3 Professor, Civil Engineering Department, Indian Institute of Technology, Hauz Khas, New Delhi-110 016, India, email- [email protected]

7

RI PT

2

Simulating ozone concentrations using precursor emission inventories in Delhi –National Capital Region of India

1

Abstract

This study simulates ground level ozone concentrations in a heavily populated and polluted National Capital Region (NCR- Delhi) in India. Multi-sectoral emission inventories of ozone precursors are prepared at a high resolution of 4x4 km² for the whole region covering the capital city of Delhi along with other surrounding towns and rural regions in NCR. Emission inventories show that transport sector accounts for 55% of the total NOx emissions, followed by power plants (23%) and diesel generator sets (7%). In NMVOC inventories, transport sector again accounts for 33%, followed by evaporative emissions released from solvent use and fuel handling activities (30%), and agricultural residue burning (28%). Refuse burning contributes to 73% of CO emissions mainly due to incomplete combustion, followed by agricultural residue burning (14%). These emissions are spatially and temporally distributed across the study domain and are fed into the WRF-CMAQ models to predict ozone concentrations for the year 2012. Model validations are carried out with the observed values at different monitoring stations in Delhi. The performance of the models over various metrics used for evaluation was found to be satisfactory. Summers and post-monsoon seasons were better simulated than monsoon and winter seasons. Simulations have shown higher concentrations of ozone formation during summers and lesser during winters and monsoon seasons, mainly due to varying solar radiation affecting photo-chemical activities. Ozone concentrations are observed lower at those locations where NOx emissions are higher, and concentrations increase close to the boundary of study domain when compared to the center of Delhi city. Downwind regions to Delhi are influenced by the ozone formed due to plume of precursor emissions released from Delhi. Considering significant background contributions, regional scale controls are required for reducing ozone in NCR.

28

Key words: Ozone, NCR, WRF-CMAQ modelling, air quality

29

1.

30 31 32 33 34

Globally, ground level ozone is a pollutant of concern (TRS, 2008) and is now being realized as an emerging air pollution issue in India also (Kumar et al., 2012; Ghude et al., 2014). Limited monitoring results show significant violations of prescribed standards in Indian cities (CPCB, 2015). Ozone concentrations are generally found to be lower in the city centers due to its reactions with primary nitric oxide (NO) released from vehicular sources (Sillman, 2003), and hence, could be higher outside the city

AC C

EP

TE D

M AN U

SC

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Introduction

∗Corresponding author :

Email: [email protected] (Mukesh Khare) Tel: 011-26591212, Fax: 011-24682144

1

ACCEPTED MANUSCRIPT

limits. Ground level ozone is widely known for its impacts on human health (WHO, 2008) and agriculture (Hayes et al., 2009, 2010). Burney and Ramanathan (2014) have estimated a loss of up to 36% in wheat crop yields in India due to ozone pollution. Sharma et al. (2016) has pointed out that measures taken at the urban centres in India can help in improving regional scale ozone pollution in the downwind regions. Urban centers in India have significant sources of emissions of NOx and VOCs, which are two main precursors for ground level ozone formation. Inventorisation of precursor emissions and air quality simulations studies are required for understanding the formation and transport of ozone to draft strategies for its control in these regions in India.

65

2.

Material and methods

66

2.1

Study Domain

67 68 69 70 71 72 73 74 75

NCR in India is the domain chosen for this study, which is the largest urban agglomeration in India and is known for its severely deteriorated air quality. The region accommodated a population of over 47 million in 2011 (RGCC, 2011). Other than the city of Delhi, there are a number of districts from three neighboring states (Haryana, Uttar Pradesh and Rajasthan) which fall within the NCR. Figure 1 shows overall study domain with key locations. For the purpose of emission inventorisation and air quality modeling, domain is divided into grids of 4x4 km² using geographic information system (GIS). There are 60 grids in x-direction (240 kms) and 72 grids in y–direction (288 km²). Population, registered vehicles, industrial fuel consumption, and other features of the important sub-regions in the NCR are presented in Table 1.

RI PT

35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

AC C

EP

TE D

M AN U

SC

National capital region (NCR) of Delhi is a critically polluted region and ozone concentrations are found to be significantly higher than the standards (CPCB, 2014; Sharma et al., 2013). The region is known for high population densities due to resident population and large inflow of international and domestic tourists. The city of Delhi, which is in the middle of the NCR, has been termed as one of the most polluted cities in the world (Cheng et al., 2016; WHO, 2015). While there are studies to assess and control particulate matter (PM) pollution in the city, there is no study targeted specifically to simulate ozone concentrations in Delhi and surrounding regions as a whole. There were efforts made in past to report emission inventories for the Delhi region (Gargava et al., 2014; Gurjar et al., 2004; Guttikunda et al., 2013; Sahu et al., 2011; Goyal et al., 2013, Marrapu et al., 2014), but limited work has been carried out in estimating ozone precursor emissions for the entire NCR. Considering chemical reactivity and transport of ozone, it is important to carry out the assessment for entire NCR. This study presents the first high resolution (4x4 km²) gridded emission inventories of the main precursors of ozone (oxides of nitrogen NOx, non-methane volatile organic compounds (NMVOCs), carbon monoxide (CO) for the entire NCR for the year 2012. The emission inventories are fed into a chemical transport air quality model to assess the formation and transport of ground level ozone in NCR. Performance of the model is evaluated by comparing the modelling results with observed ozone concentrations. This study will be useful in establishing the emission inventories of the ozone precursors and setting up a validated model for assessment of ozone in the heavily polluted NCR. The study also puts forward a general approach to carry out assessment of ground level ozone in similar urban settings. The results of this study can be used for carrying out source apportionment, evaluation of impacts, and formation of ozone management plan for the region.

76 2

ACCEPTED MANUSCRIPT

Table 1 Key features of sub-regions in the NCR

Region

Populati

Registered

Coal consumption

Coal

Gas

on

vehicles

(Industrial sector

consumption

consumption-

(million)

(million)

excluding brick kilns)

(Power

power sector

(kt)

sector) (kt)

(million m³)

-

-

-

-

-

-

-

-

8.46

4901

1276

47.83

-

687

15.34

8855

1140

39.34

-

-

0.66

17.68

-

-

0.46

186.07

-

-

0.24

480.57

7824

-

RI PT

77

Alwar

0.45

0.21

17.68

Bagpat

0.15

0.09

0.03

Jhajjar(Bahadurgarh)

0.17

Bulandshahar

0.38

0.27

Delhi

16.75

7.77

Faridabad

1.54

0.79

Gautam Budh Nagar

0.75

0.64

Ghaziabad

2.51

0.73

Gurgaon

0.90

Merrut

1.43

Panipat

0.44

TE D

M AN U

3.79

EP

Rewari

0.14

0.19

8.11

-

-

0.37

0.22

-

-

-

0.25

-

-

-

-

-

-

824.9

21580

3103

AC C

Rohtak Sonepat

-

SC

Urban areas

0.31

Rest rural areas

17.2

Total

43.49

12.52

78

Data source : RGCC (2011), NCRPB (2015), MoSPI (2010), DESAH (2013)

79 80 81 82

Delhi, being the capital city accomodates a huge popualtion base of about 16.8 million. The vehicular popualtion in the city has grown from about 3 million in 1998 to more than 7.7 million in 2012, with 64% of them being two-wheelers. Other than vehicles registered in the Delhi city, there is a large movement of vehicles from its surrounding towns like Gurgaon, Faridabad, Sonepat, Ghaziabad, and Gautam Budh 3

ACCEPTED MANUSCRIPT

Nagar. Delhi also has few power plants based on coal and gaseous fuels of about 840 MW and 1600 MW capacities, respectively. Other than these, there are power plants based on coal and gaseous fuels in the surrounding towns. Though, less in Delhi, there are frequent power cuts in other parts of NCR which leads to commissioning of standby power sources like diesel generators. There are not many polluting industries in Delhi. Most of these have been shifted outside Delhi municipal limit (DPCC, 2011). However, there is significant coal consumption in neighbouring districts of Panipat, Merrut, Faridabad, and Ghaziabad. With limited standards of NOx, NMVOCs and CO, emissions are released uncontrolled from many of these sources.

RI PT

83 84 85 86 87 88 89 90 91 92

SC

93 5

94 95 7

96 3

97

1

4

98

103 104 105 106 107 108

EP

102

8

AC C

101

2

TE D

99 100

M AN U

6

109 110

Figure 1 Study domain covering the NCR

111 112

Ozone monitoring stations: 1: ITO, 2: RK Puram, 3: Punjabi bagh, 4: IGI Airport, 5: Super regional site (outside NCR), 6: Rural site in NCR, 7: Urban site in NCR, 8: Downwind site to Delhi 4

ACCEPTED MANUSCRIPT

113

2.2

114 115 116 117 118 119

The overall approach followed in this study is presented in Figure 2. Emission inventories are prepared for the main precursors of ozone for entire NCR at a resolution of 4x4 km². Metereological model runs are performed to generate 3-dimensional metereological fields over the study domain, which are input to the chemical transport model along with emission inventories. Air quality simulaiton runs are performed for the year 2012 and the concentrations of simulated ozone in the study domain are compared with actual observations obtained from existing monitoring network for evaluating the performance of the model.

RI PT

Methodology

120

EP

TE D

M AN U

SC

121

138

123

124 125 126 127 128

129 130 131 132 133 134

AC C

137

122

135 136

Figure 2 Overall approach in this study

139

5

ACCEPTED MANUSCRIPT

140

2.2.1 Emission inventorisation

141 142

Emission inventories of ozone precursors like NOx, NMVOCs and CO are prepared using the standard methodology described in Klimont et al. (2002) (Eq. 1). =∑ ∑ ∑

,, .

1−

,, .

, , .

.

,, ,

………………………..(1)

RI PT

143

where, k is region, l is sector, m is fuel or activity type, and n is abatement technology, E is emissions of ozone precursor pollutants (NOx, NMVOC, CO) (kt); A is the activity rate; ef is the unabated emission factor (kt per unit of activity); η, the pollutant removal efficiency (%/100); X, is the actual implementation of abatement technology n (%/100); and ∑X=1

148 149

Major sources identified in NCR and considered for inventorisation of ozone precursors are described in Table 2.

150

Table 2 Sources inventorised in this study

M AN U

SC

144 145 146 147

S.No

Sector

Description

1

Transport

Two-wheelers, cars, three-wheelers, light commercial vehicles (LCVs), buses, heavy duty trucks, tractors

Industries

Coal and oil consuming industries

3

Power plants

Coal and gas consuming power utilities

4

Diesel

DG sets used as a standby source for electricity

generators(DG)

Open agricultural Waste residues from crops like wheat, rice, maize etc. burning

7

Refuse burning

Refuse burnt for volume reduction and heating purposes

AC C

6

EP

5

TE D

2

Evaporative

Paints, printing, fuel handling, personal product use etc.

sources

8

Residential

Fuel use for cooking

151 152

6

ACCEPTED MANUSCRIPT

In our previous work (Sharma et al., 2016), an emission inventory for whole of India was prepared at a resolution of 36x36 km² based on secondary data collected for various sectors. In the present paper, we have prepared a new and detailed emission inventory of NCR, at a much high resolution of 4x4 km². This was carried out by primary surveys of traffic counts and parking lots, especially for transport sector in Delhi. Many of the sources like brick kilns and other industries which were equally distributed over a region in the India-scale work (Sharma et al., 2016) have now been marked at the specific locations in NCR to improve the spatial representation of emissions. Some of the sources like refuse burning which were not considered in the earlier study have now been taken into account. All these steps helped to setup a high resolution validated model for assessment of ozone in a highly varying urban-rural landuse in NCR.

163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178

In this study, activity data for different sectors, is collected from published literature. For some sectors like transport, traffic counts are derived from primary surveys, and vehicular fleet technology mix (based on vintages, fuel types, engine types) based on parking lot surveys. Information on actual on-road traffic counts and technological mix have been used to estimate energy consumption in road transport sector in Delhi. The data on traffic counts were retrieved from survey carried out by The Energy and Resources Institute (TERI), New Delhi, which had vehicle count results at 25 locations in the study domain representing different land use categories such as residential (high, medium and low population densities), industrial, commercial, and mixed. All three types of road categories including arterial, sub-arterial and local are surveyed for number and types of vehicles passing through in a day. Data on vintage, fuel and technology wise distribution of vehicles in Delhi was also collected from surveys conducted by TERI. The overall gasoline and diesel consumed in transport sector in Delhi is estimated to be 1593 and 1710 million liters annually, respectively. The actual consumption of fuels obtained from the oil companies is found to be in close vicinity of these estimates. For rest of the NCR, the fuel consumption estimates are derived using number of registered vehicles in different districts. Survey data generated in the functional plan on Transport for NCR (NCRPB, 2010) is used to assess inter- and intra- city traffic movements in the NCR. This has also been used to assess traffic movement over highways in the NCR.

179 180 181 182 183 184 185 186 187

Data on fuel use in industries in the districts falling under NCR is taken from (MoSPI, 2011). The coal consumption in industries is found to be about 0.8 million tonnes, excluding fuel used in brick manufacturing and power plants. Google Earth (https://earth.google.com/) is used to count and mark the brick kilns in the entire NCR. It has been observed that there are about 2000 kilns which produce about 25 kt of bricks on a daily basis in the region. There are in all 5 coal based and 4 gas based power plants in the region, having an overall installed capacity of about 7000 MW. CEA (2012a) has reported a consumption of 21 million tonnes of coal consumption in power plants of NCR during the year 2011-12. The specific coal consumption in power plants varied between 0.80-0.88 kg/kwh. Gas consumption of about 8.5 MMSCMD (million m³ per day) is reported in gas based power stations (CEA, 2012b).

188 189 190 191 192 193 194

Energy consumption in the residential sector is estimated using district-wise population from RGCC (2011). Further, per capita consumption rates of different fuels used for cooking (biomass, LPG, kerosene etc.) are adopted from NSSO (2012). It is estimated that about 1 million tonnes of firewood is burnt in NCR for cooking purposes with 90% consumption in rural areas. Other than this, 0.7 million tonnes of LPG consumption is also estimated. The NCR except Delhi is known for frequent power cuts and diesel used in generator sets for standby power is an important source of NOx emissions. The diesel consumption is estimated based on per capita diesel use reported in NSSO (2012) for different states

AC C

EP

TE D

M AN U

SC

RI PT

153 154 155 156 157 158 159 160 161 162

7

ACCEPTED MANUSCRIPT

falling in NCR. District-wise population is used to derive the diesel consumption in DG sets, which is estimated to be around 175 million liters per year for the region. Burning of refuse is a common practice in the region, either for waste volume reduction or heating purposes, especially in winter season. NCRPB (2015) provides the waste generation rates for different regions in NCR, which are used to derive estimates for waste burnt in the region. With an assumption of 50% collection efficiency and 60% waste burning ratio, about 1.5 million tonnes of waste is estimated to be burnt annually. Agricultural burning of crop residues is also an important activity which contributes to emissions in the NCR. District-wise production data for different crops like wheat, rice, maize, and millet are collected and crop to waste ratio of 1.4, 2.05, 1.4, and 1 are used, respectively (Ramachandra et al., 2005).

204 205 206 207 208 209 210 211 212 213 214 215 216

The emission factors used in this study for estimating emissions of different sources are presented in Annexure A.1. ARAI (2008) is used as the indigenous emission factor database for vehicular sources. The activity data prepared for different categories, fuels, technology and vintages of vehicles is multiplied with the corresponding emission factors in ARAI (2008). Introduction of advanced vehicular emissions and fuel quality norms have also been taken into account. It is to be noted that NCR has moved to BS-IV (Euro-IV equivalent) norms from 2010. The emission factors for industrial sources are adopted from CPCB (2011). Emission factors for power plants brick manufacturing activities are adopted from GAINS Asia database (gains.iiasa.ac.at/gains/). Emission factors for agricultural residue burning and refuse burning are taken from CPCB (2011), Reddy et al. (2002), and Wiedinmyer (2014), respectively. Emission factors for different precursors for various fuel used for cooking purposes in residential sector are adopted from different local studies (Sen et al., 2014; Saud et al., 2011a; Saud et al., 2011b; Sharma et al., 2015 and Arora et al., 2013, 2014). Emission inventory of evaporative sources of NMVOC in the region is directly adopted from Sharma et al. (2015).

217 218 219 220 221 222

QGIS software has been used to allocate emissions at respective geographical locations. Transport sector emissions are allocated in the Delhi city based on road network, and on highways based on road –wise vehicle emission inventories. Industrial emissions are allocated based on district-wise fuel consumption, while bricks kiln emissions are allocated to the grids falling over actual locations of the kilns. Residential sector emissions are allocated based on population and agricultural burning emissions using district wise production datasets.

223 224 225 226 227

The NMVOC emissions were speciated using the same methodology as followed in Sharma et al. (2015), where speciated emission inventory of NMVOC has been prepared for whole of India. There is limited information about speciation profiles of NMVOCs in India, and hence, we have used sector specific species profiles developed in China (Wei et al., 2014) and SPECIATE (US EPA [http://www.epa.gov/ttnchie1/software/speciate]) database.

228

AC C

EP

TE D

M AN U

SC

RI PT

195 196 197 198 199 200 201 202 203

229

2.2.2

230 231 232 233

The WRF (Weather research forecasting model, ver 3.1.1) is the meteorological model and the CMAQ, v4.7, modeling system is used in this study to carry out simulation of ozone in varying meteorological conditions across the year (Byun and Ching, 1999; Foley et al., 2010). Figure 3 describes broad modelling framework used in this study.

Air quality simulations

8

ACCEPTED MANUSCRIPT

234 235

Met boundary conditions

236

Emissions

WPS

238

WRF

CMAQ

239

RI PT

237

Ozone concentrations

Boundary conditions

240

Figure 3 Modelling framework for air quality modelling in the study domain

242

2.2.3

243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259

The CMAQ model is based on a multi-pollutant and one atmosphere approach and is a leading air quality model used for assessment of ozone (Byun and Schere, 2006). The CMAQ modelling system consists of specific processors to take into account atmospheric processes like advection and diffusion, gas phase chemistry, plume-in-Grid processes, particle modeling and visibility, cloud processes, and photolysis rates. In gas phase chemistry module, CMAQ takes into account reactions between different ozone precursor species and meteorological factors influencing the chemistry. A number of studies have shown satisfactory performance of CMAQ to predict regional scale concentrations of a variety of pollutants. Sharma et al, 2016a have reviewed studies involving photochemical modelling and a number of studies that have successfully used CMAQ for modelling of ozone across the world. Liu and Zhang (2013) show satisfactory performance of CMAQ for predicting ozone over the southeastern U.S. Sokhi et al. (2006) and Chen et al. (2007) have demonstrated the use of CMAQ in cities like London and Beijing, respectively. Sharma et al. (2013), and Wang et al. (2015) describe use of CMAQ model in Asian cities. Satisfactory agreement between modelled and observed ground level ozone (GLO) concentrations has been reported by authors. The WRF-CMAQ combination has also been successfully used in different studies (Im et al., 2010; Cabaraban et al., 2013; Shimadera, 2011; Chatani et al., 2014) for simulation of meteorology and air quality at regional and urban scales. Hence, WRF (ver 3.1.1)-CMAQ (ver 4.7) combination has been chosen for carrying out the assessment.

260

2.2.4

261 262 263 264 265 266 267 268 269 270

WRF model was used to generate three-dimensional meteorological fields across the study domain using terrain data at a resolution of 2 min and 6-hourly average meteorological variables as inputs (NCEP, 2013; UCAR, 2010). WRF single-Moment 3-class scheme is used for micro-physics option in WRF along with RRTM scheme for Longwave Radiation, Dudhia scheme for shortwave radiation, Chen-Zhang thermal roughness length over land for surface layer, and Noah Land Surface Model: Unified NCEP/NCAR/AFWA scheme with soil temperature and moisture in four layers, for the land surface. WRF modelling results for the year 2012 are processed by the Meteorology–Chemistry Interface Processor (MCIP) for preparing CMAQ model inputs. Precursor emission inventories along with the meteorological inputs from WRF model are defined as input to the CCTM, the CMAQ chemical transport model module of CMAQ, to simulate hourly ozone concentrations at different grid points in the study

SC

241

AC C

EP

TE D

M AN U

Meteorological and air quality models- WRF and CMAQ

Modelling and validation

9

ACCEPTED MANUSCRIPT

domain for the year 2012. Simulations carried out in our previous work Sharma et al. (2016) have been used to derive boundary conditions for the study domain (NCR). BCON processor of CMAQ model has been used to derive space and time varying ozone concentrations at the boundaries of the domain.

274 275 276 277 278 279 280

The 24-hourly average modelled concentrations of ozone for the year 2012 are compared with actual observations being monitored at four ambient monitoring stations, i.e. at IGI airport, ITO, Punjabi Bagh and R.K. Puram (Figure 1). The observations at the four stations are made by the Delhi Pollution Control Committee on a continuous basis. Modelled and actual observations are compared and analysed using descriptive statistical parameters e.g. index of agreement (IA), normalized mean bias factor, normalized mean absolute error factor, mean bias, mean error and root mean square error (Yu et al., 2006). The formulae for the metrics used are presented below.

RI PT

271 272 273

i

∑

i

M i − Oi M

∑M

i

n

284

d = 1−

∑ (M i =1

∑ (M n

i =1



i

i

= 1/

287 288 289

2

!"

(

− |

−

!"

((

−

AC C

& ' = *({1/

− Oi )

− O + Oi − O

##$# = 1/

285 286

M AN U

ENMAEF =

∑O ∑M

!"

0 ≤ d ≤1

)

2

TE D

283

BNMBF = 1 −

)

|

EP

282

SC

281

)+)}

where Mi = modelled value at site/time, i ; Oi = observed value at site/time, i ; n= number of paired - = mean observed value. modelled/observed values; - = mean modelled value; O

Other than this, spatial distribution of ozone concentrations is assessed along with varying NOx/VOC ratios at different locations in the study domain and also with varying meteorological conditions.

290

10

ACCEPTED MANUSCRIPT

3.

Results and discussions

292

3.1

Emission inventory of ozone precursors in NCR

293 294 295 296 297 298 299

Based on activity data and emission factors, the emission inventory for ozone precursors-NOx, NMVOCs and CO is prepared for different sub-regions in NCR (Table 3). Total NOx emissions in the region are estimated to be about 218 kt/yr, while NMVOC emissions are estimated to be 329 kt/yr. CO emissions, which are mainly dominated by refuse burning activity are estimated to be 2949 kt/yr. NOx to NMVOC ratio in the region is 0.66. The ratio is comparatively more than the ratios earlier reported for national scale by Sharma et al. (2016). It is due to presence of significant NOx emission sources like vehicles, DG sets and power plants in the region.

300

Table 3 Region-wise ozone precursor inventory (kt/yr) for NCR NMVOC

CO

Panipat

21.9

5.4

44.4

Sonipat

1.5

2.7

19.0

Rohtak

2.8

3.0

22.9

Faridabad

14.6

15.9

104.8

Gurgaon

4.4

5.2

50.0

TE D

M AN U

NOx

2.7

20.8

5.1

9.1

130.8

3.9

6.6

91.0

1.7

2.8

35.5

0.6

1.2

10.3

3.5

3.5

102.6

88.0

107.7

520.9

Jhajjar (Bahadurgarh)

1.3

1.4

9.9

Gautam Budh Nagar

23.2

5.8

43.8

NCR-rural

43.2

155.6

1741.9

Total

218

329

2949

Ghaziabad Merrut Bulandshahar

Alwar Delhi

AC C

Bagpat

EP

2.4

Rewari

301 302

SC

Place

RI PT

291

11

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

303 304

Figure 4 a-e: Share of different sectors and regions in emission inventories of different precursors of ozone in NCR (2012)

305 12

ACCEPTED MANUSCRIPT

RI PT

Figures 4 a-e describe sectoral and region wise shares of ozone precursor emissions of NOx, NMVOC, and CO, respectively. In NCR, transport sector accounts for 55% of the total NOx emissions, followed by power plants (23%) and DG sets (7%). Transport sector again accounts for 33% of NMVOC emissions in NCR, followed by 30% emissions from evaporative emissions due to solvent use and fuel handling. 28% NMVOCs in NCR are estimated to be generated due to open burning of agricultural residues. Refuse burning activity contributes to 73% of CO emissions mainly due to incomplete combustion, followed by open agricultural residue burning activities (14%). Further, Figure 5 shows vehicle category-wise distribution of NOx and CO emissions from transport sector in Delhi region. It is evident that 64% of NOx emissions are due to heavy duty diesel vehicles (trucks, buses, and mini-buses), while two-wheelers contribute to 14%. On the other hand, gasoline driven vehicles dominate the CO and VOC emissions in Delhi.

TE D

M AN U

SC

306 307 308 309 310 311 312 313 314 315 316

317 318

324

AC C

EP

319 320 321 322 323

325

Figure 5 Vehicle category-wise distribution of NOx, VOC, and CO emissions from Delhi region

326 327 328 329

Region wise analysis shows that 40%, 33% and 18% of NOx, NMVOC and CO emissions are from Delhi region. Other urban regions in NCR cumulatively account for 40%, 20%, and 23% of emissions; and rural regions account for 20%, 47%, 59% emissions, respectively. It is to be noted that rural regions have considerably higher shares in NMVOC and CO inventories, mainly due to contributions from refuse 13

ACCEPTED MANUSCRIPT

330 331 332

burning, open biomass burning in agricultural fields, and biomass based cooking in rural households. Figure 6 describes spatial distribution of NO emissions which show higher intensities in around urban areas and national highways with heavy traffic flows.

334 335 336

SC

337

RI PT

333

338 339

M AN U

340 341 342 343

Figure 6 Spatial distribution of NO emissions in NCR

345 346 347 348 349 350 351 352 353

Table 4 presents the comparison of emission estimates generated in present study with earlier studies focusing on Delhi city. It is evident that there are some differences in emission estimates between present and earlier studies. These are mainly due to changes in source emission intensities over time. A number of power plants in Delhi have been shifted from coal to gas based in last few years. In 2010, advanced EuroIV equivalent vehicular emissions norms are introduced in NCR. Also, Euro-III equivalent norms have been introduced in whole of India from where HDVs come to the NCR. Moreover, this study is based on actual traffic count survey of vehicles in the city of Delhi and not just based on registered number of vehicles. Additionally, some of the sources like refuse burning, evaporative sources of NMVOCs etc. have not been accounted in many of the earlier studies.

356

EP

AC C

354 355

TE D

344

14

ACCEPTED MANUSCRIPT

Table 4 Comparative estimates of emissions (kt/yr) from different studies CO

Mohan et al.(2012)

210

NEERI (2011)

137

NMVOC NOx

Year

Region

85

2008

Delhi

48

168

2008

Delhi

Guttikunda et al. (2013) 1420

261

376

2010

Delhi + few towns

This study

521

108

88

2012

Delhi

This study

2949

326

218

2012

NCR

RI PT

Study

SC

357

358 3.2 Chemical transport modelling

360 361 362 363 364 365 366 367 368 369 370 371

Emissions of ozone precursors along with WRF meteorological outputs are used as input to CMAQ to carry out air quality simulations for the year 2012. WRF model runs were performed for the year 2012 and meteorological outputs were analyzed spatially and temporally. Figure 7 shows the planetary boundary layer (PBL), temperature, and wind speed values simulated using the WRF model in four different months of the year. As expected, lower temperature, wind speeds and PBL values were detected in winter months (January), and highest values were observed during summers (April-July). MCIP processor was used to convert the WRF output files into CMAQ ready inputs. The model outputs were compared with actual observations from 2010. Temperature values were well simulated showing a r2 of 0.91 and ratio of simulated to observed values within the range of 0.84-1.11. Seasonal trend of wind speeds was also well captured, however, the magnitudes were over –predicted by the model. The ratio of simulated to observed values of wind speeds varied from 1.3-3.1 for different months, with a r2 value of 0.5.

AC C

EP

TE D

M AN U

359

15

EP

372

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 7: Spatial distribution of simulated monthly average PBL, temperature and wind speed values in NCR during 2012

375 376 377 378 379 380 381 382 383 384 385 386

Figure 8 shows monthly averaged results of simulated ozone concentrations in NCR. Seasonal variations show that ozone concentrations are much higher in summers due to increased photo-chemical activity in presence of higher solar radiations. Lowest concentrations are observed during July and August months mainly on account of low photo-chemical activity due to cloud cover and lower penetration of sunlight. Monthly average ozone concentrations in the overall study domain varies between 39-57 ppb and annual average is 49 ppb; while within the city of Delhi the concentrations are lower and vary between 28-49 ppb with an annual average of 38 ppb. The chemistry between primary NO emissions and ozone (Sillman, 2003) reduces ozone concentrations within the city of Delhi. Presence of high NO emissions lead to destruction of ozone in urban regions. Also, along the highways, ozone concentrations are lower than surroundings due to depleting reactions with NO released from vehicles. Within the city of Delhi, regions in the center of city show lower concentrations of ozone due to traffic generated NO emissions. Figure 8 also depicts the significant influence of boundary conditions in the NCR. Higher ozone concentrations

AC C

373 374

16

ACCEPTED MANUSCRIPT

around the boundaries depict contributions from outside regions (through boundary conditions derived from India scale runs in Sharma et al. (2016)). Most of the times, the wind direction in NCR is from north-west directions, except in monsoons (Jul-Aug) when it has winds blowing from opposite direction also. Consequently, the north-western regions upwind to the study domain have significant influence over ozone concentrations in Delhi and its surroundings. Sensitivity studies or ISAM (integrated source apportionment module) should be used for assessing the share of boundary conditions in NCR.

393 394 395 396 397 398

Figure 8 also shows the downwind regions (mainly in the south-east direction ) to Delhi, which can be seen influenced by the plume of emissions from the Delhi city. Ozone is produced due to the urban plume of Delhi carrying ozone precursors in the downwind direction. Titration reaction of NO with ozone reduces ozone within Delhi but precursors transported to downwind directions form a plume and ozone is formed. Later in this paper, Figure 13 also shows that ozone levels in downwind direction to Delhi are higher than city, depicting formation of ozone in the plume generated from the city.

AC C

EP

TE D

M AN U

SC

RI PT

387 388 389 390 391 392

17

ACCEPTED MANUSCRIPT

399 400 401

RI PT

402 403 404

EP

TE D

M AN U

SC

405 406 407 408 409 410 411 412 413 414 415 416 417

AC C

418 419 420 421 422

423 424

Figure 8 Simulated monthly averaged concentrations (µg/m³) of ozone in NCR during 2012

18

ACCEPTED MANUSCRIPT

3.3. Performance evaluation of models

426 427 428 429 430 431 432 433

The simulated concentrations of ozone in the study domain are compared with actual ozone concentrations being monitored at four locations shown in Figure 1 (Table 5). The modelled annual average concentrations at four locations are found to have a mean bias varying from -9% to +4 % of the observed values. For daily averaged ozone concentrations, index of agreement, d, varies between 0.610.65. These differences in daily averaged modelled and observed concentrations may be attributed to daily and seasonal variations of traffic flows. Mean bias, mean error and RMSE are found to be varying between 2 to -7 µg/m³, 15-23 µg/m³ and 19-28, respectively. The performance of the model seems to be satisfactory when compared with studies reported in (Simon et al., 2012 and Sharma et al., 2016a).

434

Table 5 Performance of model in predicting ozone concentrations at different locations in study domain OBS

MOD

(Annua

%

d

Normalized

Normalized Mean

Mean

(Annual difference

mean bias

mean

Bias

error

l avg.

avg.

factor

absolute

(µg/m³) (µg/m³)

µg/m³)

µg/m³)

M AN U

Station

SC

RI PT

425

(BNMBF)

RMSE

error factor (ENMAEF)

40.2

41.7

+4%

Punjabi Bagh

50.6

46.6

-8%

IGI

80.1

72.7

RK Puram

46.0

42.8

range

of

435 436 437 438 439 440 441 442 443

0.36

+2

15

19

0.62

-0.06

0.45

-3

23

28

0.61

-0.06

0.24

-7

19

23

-7%

0.65

-0.04

0.37

-2

17

20

0.55-

-

-

-35

to 10

+35

35

0.85

AC C

model performances

0.04

-9%

EP

General

0.63

TE D

ITO

to 30-40

Figures 9 a-d describe daily variations of modelled and observed ozone concentrations at four locations in the study domain. Seasonal trends are well captured at all four locations. The modelled concentrations are found to be higher in summers and lower in winter and rainy seasons. IGI monitoring location shows observed concentrations of ozone exceeding the predicted concentrations during winters describing lower NOx emissions in the vicinity. Almost at all locations, over-prediction of ozone is estimated during months of monsoon. However, during non-monsoon, the model performs satisfactorily. 19

ACCEPTED MANUSCRIPT

448 449 450 451 452 453 454 455 456 457 458

EP

446 447

Figure 9 a-d: Comparison of daily averaged modelled ozone concentrations with actual observations at four stations in study domain

AC C

445

TE D

M AN U

SC

RI PT

444

Model performance analysis was also carried out for different seasons across various months of the year. Figure 10 shows that summer months of March-June and post monsoon months of September to November depict lower mean bias than monsoon and winter months. Monsoon seasons show higher positive mean bias, while winters depict negatives model biases. Higher positive biases during monsoons can be attributed to limitations of meteorological simulations in reproducing local rain events and chemistry during wet season. Values of mean error and RMSE were also found to be somewhat higher for monsoon and winter seasons. During winters there is decrease in boundary layer and wind speeds and ozone concentrations depend more on low-level stability, boundary conditions, diurnal emission profiles, and deposition velocities (Loon et al., 2007). On the other hand, during summers photochemistry plays the major role in defining ozone concentrations. In this study, this appears that the model has a slightly 20

ACCEPTED MANUSCRIPT

better photochemistry module and limitations in simulating winter-time ozone and meteorological phenomena. Along with this, seasonal and diurnal variations of emissions need to be refined for further improvement in the performance. However, since the model performed better during the seasons with higher concentrations of ozone (summers and post-monsoon), it can be reliably used for assessment and control of ozone during higher concentrations.

M AN U

SC

RI PT

459 460 461 462 463

TE D

464

Figure 10: Seasonal variation of model performance (averaged for four monitoring stations)

466

4. Discussions

467 468 469 470 471 472 473 474 475 476 477 478

The simulation shows that ozone concentrations in NCR exceed the prescribed standards and violating the norms frequently. Figure 11 shows the distribution of domain averaged hourly simulated ozone concentrations in NCR for the year 2012. The prescribed national standard for hourly average ozone value is 180 µg/m³ (~90 ppb), while for 8-hourly average ozone the standard is 100 µg/m³ (~50 ppb). However, these standards are set considering mainly the impact on human health only. In NCR, there are rural regions also with considerable agricultural activities (about 5.3 million tonnes of wheat and 0.9 million tonne of Rice produced annually). We find that only 19% of times, hourly ozone concentrations in NCR are below 40 ppb (~80 µg/m3), which is a threshold for impact of ozone on agricultural productivity (Ashmore, 1994). 81% of times the averaged ozone concentrations in NCR are above 40 ppb; pointing towards significant damage in crop productivities. Moreover, this threshold does not mean that there are no impacts of ozone below this level (WHO, 2000). The domain averaged ozone concentrations lie between 40-60 ppb for more than 64% times in a year.

AC C

EP

465

21

ACCEPTED MANUSCRIPT

40 36.5344

27.7701

20

RI PT

Percent

30

17.8165

SC

14.6613

10

1.61512

0 20

30

40

M AN U

1.6026

50

60

70

0

80

Ozone concentrations (ppb) 479

Figure 11 Distribution of hourly simulated domain averaged ozone concentrations during 2012 in NCR

481 482 483 484 485 486 487

Figure 12 describes average diurnal variation of domain averaged simulated ozone concentrations. It clearly describes increase in ozone concentrations during daytime due to enhanced ozone formation in presence of sunlight. Meteorological variables also lead to significant changes in seasonal ozone concentrations. Domain averaged ozone concentrations remain 106 µg/m³ during summers; 92 µg/m³ in winters; 82 µg/m³ in monsoons; and 98 µg/m³ in post-monsoon season. Further, 98% grid locations in the study domain show ozone concentrations exceeding the threshold of 40 ppb (~80 µg/m³), confirming the severity of the ozone pollution issues in the Delhi - NCR.

AC C

EP

TE D

480

22

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

488

Figure 12 Average diurnal variation of simulated domain averaged ozone concentrations during 2012 in NCR

491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509

In the middle of the Delhi city, the ozone concentrations are lower than its surroundings. It is due to reaction between ozone and NOx where NO released from different sources reacts with ozone and reduces its concentrations in the city limits. Figure 13 shows monthly average simulated ozone concentrations in different areas within the study domain, i.e. a) super regional (Site 5: upwind of NCR influenced by boundary conditions); b) NCR-rural area (Site 6: upwind of Delhi) ; c) a NCR- urban region (Site 7: upwind of Delhi); d) Site 8: downwind of Delhi city; and e) Site 1: traffic location in Delhi city. The super-regional site (Site 5) shows highest concentrations of ozone which is deeply influenced by the boundary conditions. NCR- Rural site (Site-7) shows a slight reduction (~1%)in concentrations in comparison to the super regional site. However, ozone decreases considerably (~13%) at one of the urban sites in NCR (Site-6) i.e. the city of Rohtak , which is upwind to the city of Delhi. This is mainly due to titration reactions of ozone with local NO emissions in the Rohtak city . At one of the traffic intersections i.e. ITO in Delhi city, a drastic decrease in ozone concentrations has been observed owing to titration of ozone with NO released from vehicles around the traffic intersection. It has been observed that ozone concentrations at traffic intersection are 39% lower than outskirts of Delhi city; and are about 51% lower than at the super-regional site. The spatial gradient shows a decrease in ozone concentrations from the boundary of the study domain towards the center of the city in the direction of the wind, mainly due to titration of ozone with increasing NO emissions towards the centre of Delhi city. However, Site-8 in the downwind of Delhi shows higher values of ozone than Delhi clearly depicting the influence of ozone formed by precursors in the plume of emissions released from Delhi.

AC C

EP

TE D

489 490

510 511 23

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

512

Figure 13 Monthly average simulated ozone concentrations in the study domain

514 515 516 517 518 519 520

Figure 14 shows variations in ozone concentrations and NOx emissions at different grids in the study domain. It is evident that in presence of NOx, ozone concentrations are lower due to titration chemistry. This implies that reducing NOx emissions may initially increase ozone concentrations in Delhi. However, this should not stop the measures for control of NOx in Delhi as it has its own health and ecological impacts. Also, NOx produced in Delhi is contributing to ozone formed in downwind areas. Regional scale efforts are suggested to reduce NOx and VOC emissions which can control ozone in Delhi and the surroundings.

521 522 523 524 525

Additional research study is required to understand the interactions and sensitivities of NOx and NMVOC. Besides, source-specie (CO/NOx/NMVOC) sensitivities need to be assessed towards ozone formation. In addition, role of background ozone concentration influencing the ambient ozone concentration within the study region is also required to be understood. The validated model in this study can be used for carrying out these analysis.

EP

AC C

526

TE D

513

24

ACCEPTED MANUSCRIPT

55 R² = 0.5173 50

RI PT

45 40 35 30 25 0.0001

0.001

0.01

0.1

NOx emissions (kt/yr)

1

10

M AN U

527

SC

Ozone concentration (µ µg/m³)

60

Figure 14 Modelled ozone concentrations and NOx emissions in different grids of the study domain

529 530 531 532 533 534 535 536 537 538 539 540

A monthly analysis of ambient concentrations of ozone and its precursor species, and the key meteorological factors was carried out to understand key features of the chemistry involved in ozone formation in Delhi. Figure 15 shows the monthly averaged ozone concentrations, ratios of NO, NOx (NO+NO2), and isoprene (which is an indicator of biogenic VOC emissions), and nitrate concentrations in Delhi. There is stark variation observed in monthly averaged NO/NOx ratio, which goes down significantly during summer months (Apr-June). With rise in temperature, the reaction rates for conversion of NO to NO2 increase leading to lowering of this ratio. Nitrate formation also decreases significantly with rise in temperatures. On the other hand, with higher VOC emissions (mainly from biogenic sources) in summers, the NOx to VOC ratio goes down significantly. Figure 15 shows the increasing ratio of isoprene concentration to NOx concentrations during summers. In absolute terms, NOx concentrations are reduced in summers mainly due to dispersion caused by higher wind speeds and PBL heights. All these conditions favour higher concentrations of ozone in Delhi during summers.

541 542 543 544 545 546 547 548 549 550

The months of July and August are the ones which received 70% of the total rainfall of 320 mm during 2012. Observations show that in 2012, July and August were the cloudiest months of the year and October being the clearest. The WRF model also showed a decline of 16% in short wave radiations reaching the ground surface in comparison to summer months, and ozone concentrations also show a decline owing to attenuation of photolysis process during months of July and August. The month of July when compared with June shows interesting findings. Due to cloudiness the solar radiations reduce by 12% in July in comparison to June, while, temperatures remain almost the same. A decline of only 5% has been observed in ozone concentrations during July in comparison to June. Higher temperatures driving the ozone forming reaction rates ensured lesser decrease in ozone concentrations in comparison to decrease observed in solar radiations.

AC C

EP

TE D

528

551

25

ACCEPTED MANUSCRIPT

0.35

100

0.30 80

0.25 0.20

60

0.15

40

0.10

Concentration (µ µg/m3)

120

RI PT

Ambient concentration ratios

0.40

20

0.05 0.00

0

Isoprene/NOx

NO/Nox

O3

Nitrates

M AN U

552

SC

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure 15 Monthly variations of ratios of different precursor species and concentrations of ozone and nitrates in Delhi

555

Conclusions

556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578

This study prepares a high resolution emission inventory of the ozone precursors for heavily polluted Delhi NCR. Inventories show high contributions from transport sector, power plants and DG sets in NOx emissions; while NMVOC emissions are dominated by transport, evaporative sources, and agricultural open burning. Emissions are fed into the air quality model to predict ozone concentrations, which are satisfactorily validated with actual observations. Analysis shows that ozone concentrations are lower within urban regions of Delhi NCR due to presence of high NO emissions which destructs ozone. Seasonal variations of ozone describe that concentrations are maximum during summer months and were reduced during monsoon and winter seasons. This is primarily due to varying solar radiations which subsequently affect photo-chemical activity. The performance evaluation of model shows its satisfactory performance in predicting ozone in the high emission intensive study domain. Further analysis describes that ozone concentrations decreases significantly from boundary of the study domain to the center of Delhi city. This is mainly due to titration reaction of NO with ozone which depletes the concentrations in the city centers. Overall, most areas in Delhi NCR are having high concentrations range of more than 40 ppb of ozone on an annual basis. It is also observed that ozone concentrations increase with decreasing NOx emissions. This means that measures for control of NOx emissions may actually increase ozone concentrations in the urban region. However, this does not mean that NOx should not be controlled as it is having its own health impacts along with impacts caused by nitrates in particulate form. Moreover, NOx controls in Delhi can reduce the ozone formed downwind of the city, however, regional scale controls of both NOx and VOCs are required for effective reduction of ozone in Delhi city. This also points to the need for assessment of varying sensitivities of reducing or increasing precursor emissions over the ozone concentrations in the region. Sharma et al. (2016) and this study points out to significant contributions from long range transport in India and NCR; this also needs further assessment and control for reducing ozone in specific regions in India like NCR. This study clearly demonstrates that chemical transport

AC C

EP

TE D

553 554

26

ACCEPTED MANUSCRIPT

579 580

models predict ozone concentrations satisfactorily. These models can also be used for scenario development and policy analysis for evaluating interventions for control of ozone.

581

AC C

EP

TE D

M AN U

SC

RI PT

582

27

ACCEPTED MANUSCRIPT

References

584 585

ARAI, 2008. Air Quality Monitoring Project-Indian Clean Air Programme (ICAP). Draft report on Emission Factor development for Indian Vehicles. The Automotive Research Association of India, Pune.

586 587

Arora P., Jain S., Sachdeva K., 2013. Physical characterization of particulate matter emitted from wood combustion in improved and traditional cookstoves. Energy for Sustainable Development 17, 497–503.

588 589

Arora P., Jain S., Sachdeva K., 2014. Laboratory based assessment of cookstove performance using energy and emission parameters for North Indian cooking cycle. Biomass & Bioenergy 69, 211–21.

590 591

Ashmore M.R.,Wilson R.B., (ed.) 1994. Critical levels for air pollutants in Europe, London Department of the Environment.

592 593

Burney J. and Ramanathan V., 2014. Recent climate and air pollution impacts on Indian agriculture, PNAS 111 (46), 16319-16324.

594 595

Byun D.W., Ching J.K.S., (eds.) 1999. Science algorithms of the EPA Models-3 community multi-scale air quality (CMAQ) modeling system. NERL, Research Triangle Park, NC EPA/ 600/R-99/030.

596 597 598

Byun D., Schere K.L., 2006. Review of the governing equations, computational algorithms, and other components of the Models-3 Community Multiscale Air Quality (CMAQ) modeling system, Applied Mechanics Reviews 55, 51–77.

599 600 601

Cabaraban M.T.I., Kroll C.N., Hirabayashi S., Nowak D.J., 2013. Modeling of air pollutant removal by dry deposition to urban trees using a WRF/CMAQ/i-Tree Eco coupled system, Environmental Pollution 176, 123-133.

602 603 604

CEA, 2012a. Performance Review 2011-12, URL:http://www.cea.nic.in/reports/yearly/thermal_perfm_review_rep/1112/complete_1112.pdf, Central electricity authority, New Delhi.

605 606

CEA, 2012b. http://www.cea.nic.in/reports/monthly/fuel_sup_consm_rep/gas%20based/april12.pdf , Central electricity authority, New Delhi.

607 608 609 610

Chatani S., Amann M., Goel A., Hao J., Klimont Z., Kumar A., Mishra A., Sharma S., Wang S.X., Wang Y.X., & Zhao B., 2014. Photochemical roles of rapid economic growth and potential abatement strategies on tropospheric ozone over South and East Asia in 2030, Atmospheric Chemistry and Physics 14, 9259– 9277.

611 612 613

Chen D.S., Cheng S.Y., Liu L., Chen T., Guo X.R., 2007. An integrated MM5–CMAQ modeling approach for assessing trans-boundary PM10 contribution to the host city of 2008 Olympic summer games-Beijing, China, Atmospheric Environment 41(6), 1237-1250.

614 615 616

Cheng Z., Luo L., Wang S.,Wang Y., Sharma S., Shimadera H., Wang X., Bressi M., de Miranda R.M., Jiang J., Zhou W., Fajardo O.,Yan N., Hao J., 2016. Status and characteristics of ambient PM2.5 pollution in global megacities, Environment International 89–90, 212–221.

617 618

CPCB, 2014. National ambient air quality status & trends – 2012, Central Pollution Control Board, New Delhi.

619 620

CPCB, 2015. Air quality database, URL : http://www.cpcb.gov.in/CAAQM/mapPage/frmindiamap.aspx, Central Pollution Control Board, New Delhi.

AC C

EP

TE D

M AN U

SC

RI PT

583

28

ACCEPTED MANUSCRIPT

DESAH, 2013. Statistical Abstract Haryana 2011-12, Department of Economic And Statistical Analysis Haryana, Government of Haryana.

623 624 625

DPCC, 2011. Action plan abatement of pollution in critically polluted area of Najafgarh drain basin including Okhla, Naraina, Anand Parbat and Wazirpur Industrial Areas, Delhi Pollution Control Committee, New Delhi.

626 627

Gargava P., Judith C.C., John G.W., Douglas L., 2014. A Speciated PM10 Emission Inventory for Delhi, India, Aerosol and Air Quality Research 14, 1515–1526.

628 629 630

Ghude S. D., Jena C., Chate D.M., Beig G., Pfister G., Kumar R. & Ramanathan V., 2014. Reductions in India’s crop yield due to ozone, Geophysical Research Letters 41, 5685–5691, doi:10.1002/2014GL060930.

631 632

Goyal P., Mishra D. and Kumar A., 2013. Vehicular emission inventory of criteria pollutants in Delhi. Springer Plus 2:216.

633 634

Gurjar B.R., Van A. J.A., Lelieveld J. and Mohan M., 2004. Emission estimates and trends (1990– 2000) for megacity Delhi and implications. Atmospheric Environment 38,5663–5681.

635 636

Guttikunda S.K., Calori G., 2013. A GIS based emissions inventory at 1 km x 1 km spatial resolution for air pollution analysis in Delhi, India. Atmospheric Environment 67, 101–111.

637 638 639

Hayes F., Mills G., & Ashmore M., 2009. Effects of ozone on inter- and intra-species competition and photosynthesis in mesocosms of Lolium perenne and Trifolium repens. Environmental Pollution 157(1), 208-214.

640 641 642

Im U., Markakis K., Unal A., Kindap T., Poupkou A., Incecik S., Yeniguna O., Melas D., Theodosi C., Mihalopoulos N., 2010. Study of a winter PM episode in Istanbul using the high resolution WRF/CMAQ modeling system, Atmospheric Environment 44(26), 3085–3094.

643 644

Klimont Z., Streets D.G., Gupta S., Cofala J., Fu L., Ichikawa Y., 2002. Anthropogenic emissions of nonmethane volatile organic compounds in China. Atmospheric Environment 36 (8), 1309-1322.

645 646 647 648

Kumar R., Naja M., Pfister G.G., Barth M.C., Wiedinmyer C. and Brasseur G. P., 2012. Simulations over South Asia using the Weather Research and Forecasting model with Chemistry (WRFChem): chemistry evaluation and initial results, Geoscientific Model Development 5, 619–648, doi:10.5194/gmd-5-6192012.

649 650 651

Liu, X.-H., Zhang Y., 2013. Understanding of the Formation Mechanisms of Ozone and Particulate Matter at a fine scaleover the Southeastern U.S.: Process Analyses and Responses to Future-Year Emissions, Atmospheric Environment 74, 259-276.

652 653 654

Loon M.J.W., Robert V., Martijn S., Wind P., 2007. Evaluation of long-term ozone simulations from seven regional air quality models and their ensemble, Atmospheric Environment 41(10), 2083-2097, DOI: 10.1016/j.atmosenv.10.073.

655 656 657

Marrapu P., Cheng Y., Beig G., Sahu S., Srinivas R., and Carmichael G.R., 2014. Air quality in Delhi during the Commonwealth Games, Atmos. Chem. Phys., 14, 10619– 10630, doi:10.5194/acp-14-106192014.

AC C

EP

TE D

M AN U

SC

RI PT

621 622

658 29

ACCEPTED MANUSCRIPT

Mohan M., Shweta B., Preeti G. and Pallavi M., 2012. Emission Inventory of Air Pollutants and Trend Analysis Based on Various Regulatory Measures Over Megacity Delhi, Environmental Sciences » "Air Quality - New Perspective", book edited by Gustavo Lopez Badilla, Benjamin Valdez and Michael Schorr, ISBN 978-953-51-0674-6.

663 664

MoSPI, 2011. Data base on district-wise fuel consumption in industries, Annual survey of industries Ministry of Statistics and Programme Implementation, Govt. of India, New Delhi.

665 666 667

NCEP, 2013. National Weather Service, NOAA, U.S. Department of Commerce (2000), NCEP FNL Operational Model Global Tropospheric Analyses, continuing from July 1999, http://rda.ucar.edu/datasets/ds083.2; 2013.

668 669

NCRPB, 2015. Regional Plan 2021, National Capital Region Planning Board, Ministry of Urban Development, Government of India, New Delhi.

670 671

NCRPB, 2010. Functional Plan on Transport for National Capital Region-2032, National Capital Region Planning Board, Ministry of Urban Development, Government of India.

672 673

NEERI, 2011. Air Quality Monitoring, Emission Inventory & Source Apportionment Studies for Delhi, NEERI, Nagpur.

674 675 676

NSSO, 2012. Household Consumption of Various Goods and Services in India NSS 66th Round, Ministry of Statistics and Programme Implementation, National Statistical Organisation, National Sample Survey Office.

677 678 679

Ramachandra T.V. and Kamakshi G., 2005. Bioresource Potential of Karnataka, [Talukwise inventory with management options], Energy & Wetlands Research Group, TECHNICAL REPORT NO: 109, Centre for Ecological Sciences, Indian Institute of Science, Bangalore-560012.

680 681

RGCC, 2011. Census of India – 2011, Registrar General and Census Commissioner, Govt. of India, New Delhi.

682 683

Sahu S.K., Beig G., Parkhi N.S., 2011. Emission inventory of anthropogenic PM2.5 and PM10 in Delhi during Commonwealth Games 2010. Atmospheric Environment 45, 6180-6190.

684 685 686

Saud T., Mandal T.K., Gadi R., Singh D. P., Sharma S.K., Saxena M. and Mukherjee A., 2011a. Emission estimates of particulate matter (PM) and trace gases (SO2, NO and NO2) from biomass fuels used in rural sector of Indo-Gangetic Plain, India. Atmospheric Environment 45, 5913–23.

687 688

Saud T., Singh D. P., Mandal T. K., et al. 2011b. Spatial distribution of biomass consumption as energy in rural areas of the Indo-Gangetic plain. Biomass & Bioenergy 35, 932–41.

689 690

Sen A., Mandal T. K., Sharma S. K., et al. 2014. Chemical properties of emission from biomass fuels used in the rural sector of the western region of India. Atmospheric Environment 99,411–424.

691 692

Shimadera H., Kondo A., Shrestha K. L., Kaga A., Inoue Y., 2011. Annual sulfur deposition through fog, wet and dry deposition in the Kinki Region of Japan, Atmospheric Environment 45(35), 6299-6308.

693 694

Sillman S., 2003. Tropospheric O3 and photochemical smog. in B. Sherwood Lollar, ed., Treatise on Geochemistry, Vol. 9: Environmental Geochemistry, Ch. 11, Elsevier, 2003.

695 696

Sharma S., Satoru C., Richa M., Anju G., Atul K., 2016. Sensitivity analysis of ground level Ozone in India using WRF-CMAQ models, Atmospheric Environment 131, 29-40.

AC C

EP

TE D

M AN U

SC

RI PT

659 660 661 662

30

ACCEPTED MANUSCRIPT

Sharma S., Sharma P., Khare M., 2016a. Review of studies on photo-chemical modeling of Ozone, (Unpublished)

699 700

Simon H., Kirk R.B., Sharon P., 2012. Compilation and interpretation of photochemical model performance statistics published between 2006 and 2012, Atmospheric Environment 61, 124-139.

701 702 703

Sokhi R.S., José R.S., Kitwiroon N., Fragkou E., Pérez J.L, Middleton D.R., 2006. Prediction of O3 levels in London using the MM5–CMAQ modelling system , Environmental Modelling & Software 21 (4), 566-576.

704 705

TRS, 2008. Ground-level ozone in the 21st century: future trends, impacts and policy implications, RS Policy document 15/08, Issued: October 2008 RS1276, The Royal Society, London.

706 707

UCAR, 2010. http://www.mmm.ucar.edu/wrf/src/wps_files/geog_v3.1.tar.gz , University Corporation of Atmospheric Research.

708 709

Wei W., Wang S., Hao J., Cheng S., 2014. Trends of chemical speciation profiles of anthropogenic volatile organic compounds emissions in China, 2005–2020 ,Environ. Sci. Eng. 8(1): 27–41

710

WHO, 2000. Air quality guidelines for Europe; second edition, World Health Organization, Geneva.

711 712 713

Wiedinmyer C.,Yokelson R. and Gullett B.K., 2014. Global Emissions of Trace Gases, Particulate Matter, and Hazardous Air Pollutants from Open Burning of Domestic Waste, Environmental Science & Technology 48(16), 9523–9530.

714 715

Yu S., Eder B., Dennis S.H., Schwartz S., 2006. New unbiased symmetric metrics for evaluation of air quality models. Atmospheric Science Letters 7, 26-34.

AC C

EP

TE D

M AN U

SC

RI PT

697 698

31

ACCEPTED MANUSCRIPT

716

Annexure A.1 Emission factors used in the study

717

Road transport (g/km) Engine size

Vintage

CO

HC

NOx

LCV Diesel

<3000cc

1991-96

3.07

2.28

3.03

LCV Diesel

<3000cc

1996-2000

LCV Diesel

>3000cc

Post 2000

LCV Diesel

>3000cc

Post 2000 Post 2010

3.00

1.28

2.48

3.66

1.35

2.12

3.66

1.35

2.12

2.65

0.95

1.48

19.30

2.63

13.84

6.00

0.37

9.30

SC

LCV Diesel

RI PT

Vehicle type and fuel

>6000cc

1991-2000

HCV Diesel Truck

>6000cc

Post 2000

HCV Diesel Truck

>6000cc

Post 2000

6.00

0.37

8.63

Post 2010

4.34

0.26

6.04

9.88

1.09

9.73

9.88

1.09

9.73

M AN U

HCV Diesel Truck

HCV Diesel Truck TRAILER

TE D

TRACTOR HCV Diesel Bus

>6000cc

1991-96

13.06

2.40

11.24

>6000cc

1996-2000

4.48

1.46

15.25

>6000cc

Post 2000

3.97

0.39

11.50

>6000cc

Post 2005

3.92

0.16

6.53

>6000cc

Post 2010

2.84

0.11

4.57

HCV CNG Bus

>6000cc

Post 2000

3.72

3.75

6.21

HCV CNG Bus

>6000cc

Post 2000

3.72

3.75

4.35

Passenger Cars (Petrol)

<1000cc

1991-96

4.75

0.84

0.95

Passenger Cars (Petrol)

<1000cc

1996-2000

4.53

0.66

0.75

Passenger Cars (Petrol) BS-II

<1000cc

Post 2000 (MIDC)

1.30

0.24

0.20

Passenger Cars (Petrol) BS-II

<1000cc

Post 2000 (MIDC)

0.84

0.12

0.09

HCV Diesel Bus

HCV Diesel Bus

AC C

HCV Diesel Bus

EP

HCV Diesel Bus

32

ACCEPTED MANUSCRIPT

Engine size

Vintage

CO

HC

NOx

Passenger Cars (Petrol) BS-II

<1000cc

Post 2010

0.36

0.06

0.05

Passenger Cars (Petrol) BS-I

1000-1400cc

Post 2000 (MIDC)

3.01

0.19

0.12

Passenger Cars (Petrol) BS-I

1000-1400cc

Post 2000 (MIDC)

3.01

0.19

0.12

Passenger Cars (Petrol) BS-I

1000-1400cc

Post 2000 (MIDC)

Passenger Cars (Petrol) BS-I

1000-1400cc

Post 2000 (MIDC)

Passenger Cars (Petrol) BS-III

1000-1400cc

Post 2010 (MIDC)

Passenger Cars (Petrol) BS-I

>1400cc

1991-96

Passenger Cars (Petrol)

>1400cc

1996-2000

Passenger Cars (Petrol) BS-I

>1400cc

Passenger Cars (Petrol)

RI PT

Vehicle type and fuel

0.19

0.12

1.94

0.10

0.05

1.29

0.10

0.06

2.74

0.19

0.21

2.74

0.19

0.21

Post 00 MIDC

2.74

0.19

0.21

>1400cc

Post 05 MIDC

0.84

0.12

0.09

Passenger Cars (Petrol)

>1400cc

Post 2010 MIDC

0.36

0.06

0.05

Passenger Cars (Diesel)

<1600cc

1996-2000

0.87

0.22

0.45

Passenger Cars (Diesel)

<1600cc

1996-2000

0.87

0.22

0.45

TE D

M AN U

SC

3.01

Passenger Cars (Diesel) BS-I

<1600cc

Post 2000(MIDC)

0.72

0.14

0.84

Passenger Cars (Diesel)

<1600cc

Post 2005 (MIDC)

0.06

0.08

0.28

<1600cc

Post 2010 (MIDC)

0.05

0.05

0.14

EP

Passenger Cars (Diesel)

1600-2400cc

1996-2000

0.66

0.25

0.61

Passenger Cars (Diesel)

1600-2400cc

1996-2000

0.66

0.25

0.61

Passenger Cars (Diesel)

1600-2400cc

1996-2000

0.66

0.25

0.61

Passenger Cars (Diesel)

1600-2400cc

1996-2000

0.66

0.25

0.61

Passenger Cars (Diesel)

1600-2400cc

Post 2010

0.52

0.15

0.31

Passenger Cars (CNG)

<1000cc

1996-2000

0.85

0.79

0.53

Passenger Cars (CNG)

<1000cc

1996-2000

0.85

0.79

0.53

Passenger Cars (CNG) BS-I

<1000cc

Post 2000 (MIDC)

0.06

0.46

0.74

AC C

Passenger Cars (Diesel)

33

ACCEPTED MANUSCRIPT

Engine size

Vintage

CO

HC

NOx

Passenger Cars (CNG) BS-I

<1000cc

Post 2000 (MIDC)

0.06

0.46

0.53

Passenger Cars (CNG) BS-I

<1000cc

Post 2000 (MIDC)

0.06

0.46

0.37

Passenger Cars (LPG)

1000-1400cc

1996-2000

6.78

0.85

0.50

Passenger Cars (LPG)

1000-1400cc

1996-2000

Passenger Cars (CNG) BS-I

1000-1400cc

Post 2000 (MIDC)

Passenger Cars (CNG) BS-I

1000-1400cc

Post 2000 (MIDC)

Passenger Cars (CNG) BS-I

1000-1400cc

Post 2000 (MIDC)

MUV Diesel

<3000cc

1991-96

MUV Diesel

<3000cc

MUV Diesel BS-I

RI PT

Vehicle type and fuel

0.85

0.50

0.60

0.36

0.01

0.60

0.36

0.01

0.60

0.36

0.01

2.49

1.39

1.70

1996-2000

1.38

1.39

0.65

<3000cc

Post 2000 (MIDC)

1.94

0.89

2.46

MUV Diesel

<3000cc

Post 2005 (MIDC)

0.25

0.19

0.67

MUV Diesel

<3000cc

Post 2010

0.20

0.11

0.34

Moped (2 Stroke)

< 80cc

1991-96

11.41

7.70

0.02

< 80cc

1996-2000

2.97

2.77

0.03

< 80cc

Post 2000

0.45

3.10

0.04

< 80cc

Post 2005

0.46

0.60

0.02

< 80cc

Post 2010

0.31

0.40

0.01

< 80cc

1996-2000

5.20

2.51

0.04

Scooter (2 Stroke)

< 80cc

Post 2000

2.37

2.05

0.03

Scooter (2 Stroke)

> 80cc

1991-96

6.00

3.68

0.02

Scooter (2 Stroke)

> 80cc

1996-2000

5.10

2.46

0.01

Scooter (2 Stroke)

> 80cc

Post 2000

2.76

2.16

0.03

Scooter (2 Stroke)

> 80cc

Post 2005

0.16

0.86

0.02

Scooter (2 Stroke)

> 80cc

Post 2010

0.11

0.58

0.01

Moped (2 Stroke) Moped (2 Stroke)

AC C

Scooter (2 Stroke)

M AN U

TE D

Moped (2 Stroke)

EP

Moped (2 Stroke)

SC

6.78

34

Engine size

Vintage

CO

HC

NOx

Motorcycle (2 Stroke)

< 80cc

1991-96

5.64

2.89

0.04

Motorcycle (2 Stroke)

> 80cc

1996-2000

2.96

2.44

0.05

Motorcycle (2 Stroke)

> 80cc

Post 2000

2.96

2.44

0.05

Motorcycle (2 Stroke)

> 80cc

Post 2010

Moped (4 Stroke)

< 100cc

Post 2000

Moped (4 Stroke)

< 100cc

Post 2010

Scooter (4 Stroke)

>100cc

1991-96

Scooter (4 Stroke)

>100cc

1996-2000

Scooter (4 Stroke)

>100cc

Scooter (4 Stroke)

>100cc

Scooter (4 Stroke)

>100cc

Motorcycle (4 s)

<100cc

Motorcycle (4 s)

<100cc

RI PT

Vehicle type and fuel

M AN U

ACCEPTED MANUSCRIPT

0.03

0.81

0.50

0.29

0.41

0.25

0.15

0.93

0.65

0.35

0.93

0.65

0.35

Post 2000

0.93

0.65

0.35

Post 2005

0.40

0.15

0.25

Post 2010

0.27

0.10

0.17

1991-96

3.12

0.78

0.23

1996-2000

1.58

0.74

0.30

<100cc

Post 2000

1.65

0.61

0.27

<100cc

Post 2010

0.83

0.31

0.14

100-200cc

Post 2000

1.48

0.50

0.54

100-200cc

Post 2010

0.74

0.25

0.27

>200cc

Post 2005

0.72

0.52

0.15

Motorcycle (4 s)

>200cc

Post 2010

0.48

0.35

0.10

Three Wheelers (2Stroke)

<200cc

1996-2000

3.15

6.04

0.30

Three Wheelers (2Stroke)

<200cc

Post 2000

1.37

2.53

0.20

Three Wheelers (2Stroke)

<200cc

Post 2005

1.15

1.63

0.16

Three Wheelers (2Stroke)

<200cc

Post 2010

0.77

1.09

0.11

Three Wheelers (4Stroke)

<200cc

1991-96

4.59

1.63

0.60

Motorcycle (4 s) Motorcycle (4 s)

Motorcycle (4 s)

AC C

Motorcycle (4 s)

EP

Motorcycle (4 s)

35

SC

1.23

TE D

1.49

ACCEPTED MANUSCRIPT

Engine size

Vintage

CO

HC

NOx

Three Wheelers (4Stroke)

<200cc

1996-2000

4.59

1.63

0.60

Three Wheelers (4Stroke)

<200cc

Post 2000

4.59

1.63

0.60

Three Wheelers (4Stroke)

<200cc

Post 2005

2.29

0.77

0.53

Three Wheelers (4Stroke)

<200cc

Post 2010

Three Wheeler Diesel

<500cc

1996-2000

Three Wheeler Diesel

<500cc

Post 2000

Three Wheeler Diesel

<500cc

Post 2005

Three Wheeler Diesel

<500cc

Post 2010

Three Wheeler CNG OEM 4S

<200cc

Three Wheeler CNG Retro 2S

<200cc

Three Wheeler LPG (Retrofit 2S)

<200cc

Three Wheeler LPG (Retrofit 2S)

<200cc

719

Domestic sector Kg/t Firewood 1.7 27.23 11.395

AC C

720 721

L.P.G. 2.25 30 9.57

EP

NOx CO HC

0.36

9.16

0.63

0.93

2.09

0.16

0.69

0.41

0.14

0.51

0.21

0.08

0.42

Post 2000

1.00

0.26

0.50

Post 2000

0.69

2.06

0.19

1996-2000

7.20

5.08

0.05

Post 2000

1.70

1.03

0.04

SC

0.52

M AN U

Source : ARAI, 2008

1.53

TE D

718

RI PT

Vehicle type and fuel

Kerosene 1.3 43 8.54

Industries Coal Kg/t Nox 4.8

SO2 9.75

CO 7.3

HC

Oil Kg/t NOx SO2 6.6 0.7536

722 723

36

CO 0.6

HC 0.13

ACCEPTED MANUSCRIPT

Brick

CO 10

VOC -

725 726

Power Plant Coal Kg/t Nox 2.23

CO 0.3

HC 0.3

Gas kg/MMSCM NOx 376

728

Refuse burning Kg/t NOx 3.74

CO 1453

NMVOC 22.6

729 730

DG set

M AN U

727

734 735

HC 22.9

AC C

EF(kg/t) NOx CO 0.49 58

EP

Open agricultural burning

TE D

EF (kg/kwh) NOx VOC CO 0.018848 0.001529 0.00406144 731 732 733

RI PT

EF (kg/t) NOx 0.255

37

CO 1344

SC

724

HC 88

ACCEPTED MANUSCRIPT

Highlights

RI PT

Sectoral emission inventories of ozone precursors at a high resolution for NCR. Simulated ozone concentrations in heavily populated and polluted NCR in India Satisfactory performance of the models for ozone prediction Higher concentrations of ozone during summers than in winters and monsoon seasons Ozone is found to be lower at locations of higher NOx emissions Ozone is highest close to the boundary of study area and lowest at center of city

AC C

EP

TE D

M AN U

SC

• • • • • •