Quantitative Assessment of the Emitted Criteria Pollutant in Delhi Urban Area

The compositions of the emitted criteria pollutant from the anthropogenic sources are investigated in the present study for Delhi urban area. The quantitative study of various types of the criteria pollutants, i.e., Carbon Monoxide (CO), Nitrogen Oxide (NOx) and Particulate Matter (PM) emitting from various sources is carried out. The primary emitting sources of these criteria pollutants are the vehicles, industries, power plants, domestic and dust sources. The estimated emissions depends a number of factors such as fuel consumption, technological developments, industrial activity and other activities that cause air pollutions. The analysis shows that CO, NOx and PM emissions are mainly contributed by vehicular sources. The results of the present study are found to be well compared to those reported in literature. The estimated emissions are also shown by spatial variations isopleths over 26 km × 30 km gridded area of Delhi. The concentrations of these pollutants are predicted by using the emissions and meteorological variables as the input in AERMOD, an USEPA’s dispersion model. Uncertainties in the emission estimates, measured as 95% confidence intervals, range from a low of ± 32% for NOx to a high of ± 235% for CO. Quantile-Quantile (Q-Q) plots and statistical performance have also been analyzed and it is observed that the predicted concentrations of CO, NOx and PM are in good agreement with respect to observed values. Further, in some locations the predicted concentrations are much higher than the permissible National Ambient Air Quality Standards (NAAQS) levels. These locations are ITO junction, DCE, ISBT Kashmere Gate, Dhaula Kuan, Indraprashta and Badarpur. Thus, this study shows high air pollution load in Delhi and degrading the ambient air quality creates many environmental problems.


INTRODUCTION
Air pollution ranks for the first time among the top ten most deadly diseases (GBD, 2013) in the world as both, indoor and outdoor activities generate the air pollution.The indoor activities involve cooking or smoking while outdoor pollution owes it to industry, vehicles and power plants.The activities of combustion, construction, mining, agriculture and the power plants are becoming increasingly significant as a major contributor of air pollution.According to a study over a million people in the world, the number of people dies every year due to the air pollution in India is the highest (Chauhan, 2013).However, air pollution is identified as the fifth largest killer and was responsible for 0.6 million premature deaths in India, due to an almost six times of the increment of air pollutions to that in the year 2000 (GBD, 2013).Ambient air quality is the most serious environmental concern in urban areas around the world, especially in developing countries.The development of a detailed emission inventory is an important step in an air quality management process for any urban area.Emission inventories, generally help to determine significant sources of air pollutants, establish emission trends over time, target regulatory actions, and used in air quality forecasting models.They estimate the emission rates from various air pollution sources in a specific geographical area.Therefore the main objective on the present study to estimate the reliable emission inventories to use in air quality modeling.
For the past two decades the urban agglomeration of Delhi has witnessed a phenomenal growth in terms of the urban infrastructure, vehicular growth and the industrial activities.This urban development has aggravated the air pollution problem in Delhi.The major sources of air pollutants are attributed to the vehicular traffic, domestic, industrial and thermal power plants, etc.These are releasing the harmful chemical compounds in the atmosphere of Delhi.Among them the main criteria air pollutants categorized by National Ambient Air Quality Standard (NAAQS), India are: PM, CO and NO x .Basically, these particles are coming from combustion in vehicles, power plants, wood fires and industrial processes.Experts are blaming air pollutants from the burning of fuels, industrial factories, construction sites, brick kilns and vehicular movement for the city's deteriorating environment.From emissions point of view, the anthropogenic sources are categorized namely point, area and line source.The point sources represent industries and power plants.Area sources include household/domestic cooking, bakeries and restaurants, major construction sites, crematoria, road dust and generators.The line source represents the total vehicle fleet plying over different types of road in Delhi.Among these sources, the line sources are major contributors to the pollution load in Delhi (CPCB, 2010).
According to the Delhi Statistical Handbook (DSH) 2012, the total number of registered vehicles in the Delhi region for the year 2010-2011 was nearly 7 million, i.e., more than the combined number of vehicles in another three megacities, i.e., Kolkata, Mumbai and Chennai in India (CPCB, 2008).The vehicle fleet mostly comprises of 2-wheelers (2W), personal cars (PC), 3-wheelers (3W), light commercial vehicles (LCV), heavy commercial vehicles (HCV) and buses.In Delhi, the personal mode of transportation, i.e., 2W combined with cars (four wheelers, excluding taxis) account for approximately four-fifth of the total vehicular population in Delhi (Goyal et al., 2013;Mishra and Goyal, 2014).The growing urban agglomeration is witnessing a steep rise in LPG usage, while the use of kerosene and coal is fast depleting and restricted to only slum pockets of the city.Apart from outdoor air pollution, the household usage, fuel leads to high exposure of indoor air pollutants and hence having a considerable health risk.The area source mainly aggravates CO and PM emission load in the city (DPCC, 2002).The point source mainly consists of the industries and thermal power plants.After the Supreme Court of India, a directive for the relocation/shutdown of highly polluting industries in Delhi, the remaining industrial units are scattered only in 42 industrial estates (CPCB, 2010).The total number of polluting industries in Delhi is 17,112 and out of which, 2266 units are of air polluting nature (DPCC, 2002).However, the emission levels of these industries are not much of a problem (Goyal and Sidhartha, 2003).
In addition, there are thermal plants in Delhi out of which Badarpur and Raj Ghat are coal based while Pragati and Indraprashta are gas based.These power plants are the major contributor towards the CO, NO x and PM emission load in Delhi.Currently, a series of construction activities in and around the city (domestic, metro and flyovers) are also contributing significant emissions to the PM as road dust.After introducing the compulsory use of compressed natural gas (CNG) as fuel in buses, taxis and auto-rickshaws, closed polluting industries and expanded green spaces in Delhi, air quality had significantly improved since 2001.However, in the latter half of this decade, the city's air was becoming as polluted as before.Yet, the truth is that air pollution does not get the attention from the concerned authorities.
There are different studies available over Delhi region with many drawbacks as well as none is presented gridded emissions from all types of sources in Delhi.Kansal et al. (2011) studied emissions of NO x and PM at different locations due to all fueled vehicles for the year 2004-05 and Nagpure et al. (2011) studied emissions of CO and NO x over Delhi due to only petrol vehicles for the years 1995-2005 excluding 2 wheelers.The study of Sahu et al., 2011 is based on the emissions of PM 2.5 and PM 10 for the year 2010.Further, Gurjar et al. (2004) estimated the emissions of the same pollutants due to all types of sources up to the year 2000 and however the present study, explore the total emissions from vehicles, industries, power-plants, household and road dust over the study area of 26 km × 30 km of Delhi, i.e., the central region of National Capital Region (NCR) Delhi.Thus the present study in the first study in which the gridded emissions of criteria pollutants have been discussed from four types of sources over Delhi.

Study Area
Delhi lies at an altitude of between 700 and 1000 feet and covers an area of approximately 1500 km 2 situated on the banks of Yamuna River.The chosen study area of Delhi (26 × 30 km 2 ) is shown in Fig. 1.It has a high-congested road density, covering all power plants, covering the maximum population and industrial region.ITO junction, the centrality of the study area, has been selected for analysis and validation of the estimated emissions through air quality dispersion model.It is one of the heavily congested traffic junctions in the city and is spaced at few hundred meters distance from the Bus Rapid Transit (BRT) corridor.This junction handles the maximum traffic during office hours, i.e., 8:00 to 10:00 IST and 5:00 to 7:00 IST.Two traffic streams, one from West/Central Delhi to Trans Yamuna, and the other from South Delhi to North Delhi, lead to such heavy pedestrian and vehicular demand at these locations.

Activity Data
In this study, the data are taken from different organizations to estimate the emissions from anthropogenic sources in Delhi.These organizations are the Central Pollution Control Board (CPCB), National Environmental Engineering Research Institute (NEERI), The Energy and Resources Institute (TERI), Delhi Pollution Control Committee (DPCC) and Central Road Research Institute (CRRI), New Delhi.The observed concentration data of criteria pollutants are collected from CPCB at different locations viz., ITO junction, Delhi College of Engineering (DCE) and Sirifort in Delhi.The meteorological data are collected from Indian Meteorological Department (IMD), New Delhi.The total number of registered vehicles and industries during the study period 2010 has been obtained from DSH (2012).The used emission factors for vehicles, power plants and industrial sources, etc., are taken from CPCB (2010).

Estimation of Emissions I. Vehicular Emissions
Delhi is the most-polluted city and ranked in fourth place in the world (DPCC, 2002).The vehicular sources in Delhi have been responsible for rates as high as 72% of the total amount of air pollution (Goyal et al., 2006).However the contribution of transport sector has increased day by day, but available information about total emissions are still uncertain.The total number of registered vehicles in Delhi was approximately 6.9 million for the year 2010-11 and vehicle categories were 2W (62.6%),PC (32%), goods vehicles (includes LCV and HCV 3%), 3W (1.2%) and buses (< 1%) (DSH, 2012).The emissions of CO, NO x and PM due to vehicles have been estimated for the year 2010 through USEPA's International Vehicle Emission (IVE) model.The input data in IVE model is prepared according vehicular activity, fleeting characteristics and emission factors based on local conditions.The basis of the emission prediction process of the IVE model begins with a base emission rate and a series of correction factors, which are applied to estimate the amount of pollutants from a variety of vehicle types.One can estimate the adjusted emission rate by multiplying the basic emission rate by each correction factor using Eq. ( 1) (Goyal et al., 2013).
where, Q

II. Power Plants Emissions
The electricity supply in Delhi is from two coal-fired (Rajghat Thermal Power Station and Badarpur Thermal Power Station) and two gas-fired (I.P. Gas Turbine Power Station and Pragati Power Station) power plants.The bank of River Yamuna, near central Delhi, is situated three of the plants, whereas Badarpur is located in the southeastern part of the city.The emissions from these power plants is calculated through the gross generation and fuel use with the modified formula of Gurjar et al. (2004).
Gross generation (GWh) = plant load factor (%) × capacity (GW) × 24 × days; (2) Fuel use (kt) = gross generation × fossil fuel use per GWh. (3) The activity data such as capacity of a power plant is known, plant load factor and fuel consumption of coal as well as gas are taken CEA (2010).The power plants operated hours data were collected from DPCC and emission factors data are taken from CPCB (2010).Finally, the emissions can be estimated by multiplications of fuel use with corresponding emission factors.

III. Industrial Emissions
A total of 42 industrial regions identified with air polluting potential as per the database from DPCC (2002).Information relating to consumption of fuels such as Furnace Oil, Light Diesel Oil (LDO), High Speed Diesel (HSD) and Compressed Natural Gas (CNG) by the industries in the Delhi study area is collected from different reports e.g., DPCC (2002), GNCTD (2010).All emissions are estimated for all types of industries using Eq. ( 4) (Mohan et al., 2012): where, i is fuel type and j is the year.E i,j : emission rate of industries (gm/year); C i,j : consumption of fuel per year (gm/year); EF i,j : emission factor of the particular fuel type (gm/kg).The emission factors data used for emission calculations are taken from CPCB (2010).

IV. Household Emissions
The robust increase in the population from approximately 14 millions in 2001 to 17 million in 2011 (Census, 2011) is markedly altered the emission scenario of Delhi.Household emissions contribute significantly to the total emission profile of different criteria pollutants.The fuels used for household are mainly Liquefied Petroleum Gas (LPG), kerosene, firewood, dung cake and crop waste.The LPG is the preferred option as the major population resides in urban areas and it is the most convenient, least polluting and most efficient fuel for cooking.Kerosene is normally used as a transitional or back-up fuel and firewood; dung cake and crop waste is still the choice of fuel among poorer households.The emission rates of pollutants from different fuel type are estimated using Eq. ( 5) (Mohan et al., 2012): where, E i : emission per compound (i); Fuel j : consumption of fuel per fuel type (j) ; EF i,j : emissions of compound (i) per unit of fuel (j) consumed.Emission factors are taken from CPCB (2010).

V. Dust Emissions
Dust emissions are basically from road dust, construction, mining operations and windblown dust from vacant lands.Particulate emissions from paved roads are due to direct emissions from vehicles in the form of exhaust, brake wear and tire wears emissions, and re-suspension of loose material on the road surface (Nagpure et al., 2012).The quantity of dust emissions from unpaved road varies linearly with the volume of traffic.Construction operations are a significant source of dust emissions that includes both residential and non-residential construction as well as road construction.During the summer season, along with the dust, the wind carries heavy metals to the Indo-Gangetic basin, causing severe air pollution and degradation in visibility.The dust emissions data are collected from different previous studies, e.g., Nagpure et al. (2012), Guttikunda and Calori (2009).

Category-Wise Emissions
The region-specific emissions of criteria pollutants over Delhi urban area for the year 2010 are summarized through the above described methodologies.The total emissions are estimated from all direct releases, including coal-fired and gas based power plants, industrial, transport and domestic sources.The total emissions of CO, NO x and PM are approximately 234.72 kt, 133.41 kt and 38.23 kt respectively.Emissions varied considerably from locations to locations in Delhi.The important emission locations are ITO junction, ISBT Kashmiri Gate and Nizamuddin for vehicular source emissions, Badarpur and Indraprashta locations for powerplants emission and Okhala, Shazadabagh and DCE locations for industrial source in Delhi.
For CO emissions, vehicles contributing 210.83 kt, household -19.6 kt, industries -3.34 kt and power-plants -0.94 kt for the year 2010.It can be concluded that the transport sector is the largest source of CO emissions followed by household emissions among these four.The reason for increasing CO emissions is the consistent increase in the number of vehicles (mainly 2W and PC) and domestic fuel consumption.NO x emissions in Delhi are 133.4kt for the year 2010 from all main sources.Main emission sources are transported and household.The transport sector contributes maximum 68.7% of total NO x emissions followed by domestic 17.14%, industries 10.67% and power plants 8.21%.The relatively small increase emission trend in transport sector can be explained by the implementation of CNG in vehicles and the change of traditional fuel in cooking gas.These make up for emission increases of NO x in Delhi.Again, the total PM emissions in Delhi are 38.23 kt in 2010 including 25.50 kt from road dust.It is observed that 4.80 kt PM are emitted by transport, 4.08 kt from industries, 3.32 kt from power-plants and 0.53 kt from domestic sources.Here transport sources are contributing maximum PM emissions followed by road dust.In the transport sector, freight movement via HCV and LCV are the highest contributors.All the HCV is diesel operated and most of LCV is petrol operated.Phasing out of diesel vehicles and the introduction of CNG vehicles was an effective during the 2001-2004 with an initial decrease in PM emission in Delhi but now consistent increase in the number of vehicles as well fuel consumption together led to an increase in overall PM emissions.

Spatial Emission Contributions
The Gridded emissions of CO, NO x and PM provide the distribution of average emission intensities in 2010, as shown in Fig. 2 (a-c).It is obvious that these emissions are found to be uneven, and this can be related to road density, economic development and population density.From Fig. 2 The locations in Delhi, where NO x intensities are observed maximum (more than 2000 tons) for the year 2010 and except these the distributions are approximately uniform.These locations are Indraprashta (3842 tons), Badarpur (3074) and Dhaula Kuan (2470 tons) while minimizing emissions is observed at Delhi Cantonment (0.76 tons).Indraprashta region is contained two gases fueled power plants and are emitting more NO x than coal fueled power plants.Since, Dhaula Khuan is also in the list of hot locations on traffic point of view while Delhi Cantonment is the region of the army.
Fig. 2(c) gives the spatial distribution of annual PM emissions over Delhi, for the year 2010.It is observed that PM emissions are also uneven and can be related to vehicle, power plant and industrial activities.The maximum PM emissions are observed at Badarpur (2011 tons), ITO (703 tons), Raj Ghat (665 tons) and Okhla (629 tons) and minimum again at Delhi Cantonment (1.31 tons).Since, Badarpur and Raj Ghat both are coal-based power plants with installed capacities 705 MW and 130 MW respectively, while Okhla is the combination of three industrial areas.Road dust is also contributing 13.26 tons over the study area.

Comparisons with Other Studies
Fig. 3 is showing some period specific studies for different pollutants and their comparison with the emissions of the present study.Gurjar et al. (2004) prepared the comprehensive emission inventory for megacity Delhi, over the period 1990-2000 using emission factors of Kandilkar and Ramachandran, 2000.While in present study, the emission calculated over the year 2010 using emission factors of CPCB, 2010.In comparison to Gurjar et al. (2004), CO emissions decreased approximately 53%, which are due to the Bharat Stages.The decrement in the emissions of PM is observed approximately 80%, while NO x decrement is observed minimum due to the increment of CNG running vehicles.Overall, it is also found that there is the decrement in the emissions of pollutants from 2000 to 2010.CPCB along with NEERI prepared an emission inventory for Delhi based on data of collected between the years 2006-2007.The CO emission from all sources in that study was observed     Guttikunda et al. (2013), covering all the busy traffic locations, all power plants, maximum industries and domestic sectors.The present study area is only 10% study area of Guttikunda et al. (2013) while produces approximately 49% of NO x and PM emission and 22% of CO emission.Thus, it can be concluded that central Delhi produces maximum emission and require more research over this region

Comparison with Concentrations: AERMOD
Emission inventories are based on basic assumptions, available statistical data and formulations.Thus, there are no standard methods for validation of emission inventories.In such cases, concentrations are calculated using estimated emissions with the help of meteorological variables.The predicted concentrations can be compared with observed concentrations to validate of calculating emissions.Since, the meteorological variables play only limited part in the control or elimination of air pollution and it does not affect the source strength of the pollutants emitted.The correlation matrix of daily air pollutants and meteorological variables in Delhi has been shown in Table 1.The fundamental parameter is the wind, its speed and direction, which in turn are interrelated with the vertical and horizontal temperature gradients in the atmosphere, both of large and of small scale.Essentially, the greater the wind speed with temperature, the greater is the turbulence and the more rapid and complete is the dispersion of pollutants in the atmosphere because pollutants concentrations with wind speed are in negative correlations (Table 1).It is also observed that humidity and pressure are in positive correlations with air pollutions in Delhi.
The observed air quality in Delhi has been noticed worst during the study period.The maximum values from the daily average concentrations were observed 7990 µg/m 3 for CO, 579 µg/m 3 for PM and 276 µg/m 3 for NO x (Fig. 4).It is concluded that maximum times the daily pollutant concentrations were above the prescribed NAAQS levels.As, it has been observed that more than 75% times the PM concentrations are above the NAAQS level, while CO and NO x are approximately 50% during the study.Here, the daily average concentrations of CO, NO x and PM are predicted through the AERMOD (AMS and EPA Regulatory MODEL) model, a steady-state plume model.AERMET is pre-processed by AERMOD model, which provides the meteorological variables in the required format for AERMOD.The input files in AERMET contain the hourly surface meteorological variables and the upper air sounding data twice in a day at the same locations.The pre-processor model, AERMET, calculate boundary layer parameters for AERMOD viz.frictional velocity, Monin-Obukhov length, convective velocity scale, temperature scale, mixing height, surface heat flux by using local surface characteristics in the form of surface roughness and Bowen ratio in combination  with standard meteorological observations (wind speed, wind direction, temperature and cloud cover).These parameters are then passed through an interface present in AERMOD to calculate vertical profiles of wind speed, lateral and vertical turbulent fluctuations, and the potential temperature gradient (USEPA, 2004).The hourly surface meteorological observations have been acquired from Indian Meteorological Department (IMD), Delhi and twice daily upper air soundings have been acquired from university of Wyoming's website archive to prepare the meteorological input files for AERMOD.The predicted concentrations from AERMOD are varying between 719.3 to 5511.8 µg/m 3 for CO, 14.3 to 210.79 µg/m 3 for NO x and 36.95 to 453.73 µg/m 3 for PM.It is observed that the predicted values are under-estimated in comparisons of observed values at ITO junction over the Delhi study area.It may be concluded from here that the accurate estimation of concentration contributions from also other emission sources except for this study.
The Q-Q plots have been used extensively to evaluate the results of a model in many studies and have been considered as a useful tool for characterizing the model performance (Venktaram et al., 2004).Here, Q-Q plots are used to compare the predicted and observed concentrations of CO, NO x and PM pollutants at ITO junction.In Q-Q plots, the sorted predicted concentrations are plotted against the sorted observed values (i.e., independent of time and position) to determine whether the observed and predicted concentrations datasets come from populations with a common distribution.If the two sets come from a population with the same distribution, the points should fall approximately along the 1:1 reference line.The greater the departure from this reference line, the greater the evidence for the conclusion that the two data sets have come from populations with different distributions.The values close to the 1:1 line indicate a good fit between the predictions and observation data.The values above the 1:1 solid line shows the level of over-prediction and values lying below it under-prediction.The values on dashed line shows the predicted values are half of the corresponding observed values.Figs.5(a-c) show a Q-Q plot for 24 hours average CO, NO x , PM observed concentrations and predicted concentrations respectively for the year 2010 at ITO junction.The plots of Fig. 5(a) reveal the tendency of the model towards under-prediction at the higher end of the observed concentration distributions and over-predictions at the lower end.Also, AERMOD performs well for CO pollutant as most of the quintile point's lie along the 1:1 reference line.Similarly, AERMOD also performing well for NO x pollutant as the maximum quintile points lying around 1:1 reference line (Fig. 5(b)).However, there is a consistent tendency towards under-prediction for PM pollutant estimation through AERMOD in year 2010 (Fig. 5(c)) because maximum quintiles are at the higher end of observed concentrations.Overall, the observed data are convincible with respect to predicted concentrations at ITO junction.Thus the predicted concentrations are in good agreement with the observed values.

(B) Statistical Performance of Model
The concentration predictions of CO, NO x and PM from  by the mean) of the contributing factors (Streets et al., 2003).Further, the combination of these uncertainties to estimate the uncertainty of emissions.The overall uncertainty in emissions for all types of sources are measured as 95% confidence intervals: ± 32% (NOx), ± 63% (PM) and ± 235% (CO).So for example, ± 95% confidences of NOx emissions are within ± 32% of the state value.Similarly PM and CO is within ± 63% and ± 235% of the state value respectively.

CONCLUSIONS
The present study explores total emissions from the different emission sources viz.vehicular, power plants, industries, household and road dust for the year 2010 in Delhi.The inventory shows that the total emissions of CO, NO x and PM are 234.71kt, 133.40 kt and 38.23 kt respectively over 26 km × 30 km region.The largest contributor of CO and NO x emissions are transport followed by the domestic sector, while vehicular, industries and power plants are approximately having equal contribution in PM emission in Delhi.The spatial distribution patterns of emissions showed that each of species emissions in central and southern part of Delhi is much higher than those in other parts of Delhi.The average maximum emission of CO, NO x and PM are observed at the ITO junction as 3247, 1513 and 16.3 tons for the year 2010.The maximum emissions of different pollutants are observed at different locations in Delhi viz.CO at ISBT Kashmere Gate (3357 tons), NO x at Indraprashta (3842 tons) and PM at Badarpur (2011 tons).The predicted concentrations corresponding to the estimated emissions are observed in good agreement with the observed concentrations at ITO junction.The uncertainty analysis is also validating the prepared emissions in Delhi urban area.However, unplanned urban planning and the failure to take the environment and public health into consideration are emerging as big challenges to calculate the exact emissions in Delhi.The levels of CO, NO x and PM are also found in alarming proportions in several locations that include ITO junction, DCE, Siri-Fort, Nizamuddin, Shahadara and Shahzadabad Bagh.Thus, Delhi's facade has completely changed and in less than a decade.Further, it is expected that the development of the more efficient way to calculate the emissions from all sources in Delhi to provide an advanced modelling and analysis system for environmentalists, planners and decision makers.Finally, to combat pollution in Delhi, the authority should establish and enforce air pollution standards and to set emission standards for emission sources.

APPENDIX
The statistical performances are used for the evaluation of AERMOD are given by Chang and Hanna (2004)  The normalization by ensuring that NMSE will not be biased towards models that over-predict or under-predict.Ideal value for NMSE is zero.Smaller values of NMSE denote better model performance.

Index of Agreement (IOA):
The values of agreement (IOA) compare the difference between the mean, the predicted and observed concentrations, indicating the degree of error, free for predictions.
[t]  : adjusted emission rate for each technology (start (g) or running (g/km); B[t]  : base emission rate for each technology (start (g) or running (g/km).The correction factors used in Eq. (1) are categorized into several categories as local, fuel quality and power & driving variables.The ambient temperature, humidity and altitude, etc., are used in the IVE model to calculate the emissions from vehicles.The fuel quality variables are gasoline and diesel.The power and driving variables are road grade and air conditioning usage(IVEM, 2008).The correction factors of eight parameters in Eq. (1) are taken from IVEM BERCF 2008 and IVEM Correction Factor Data 2008.The vehicular activity data have been taken same as discussed inGoyal et al. (2013).Bharat Stages IV was started in the year 2010 but it is observed that the largest ratio of personal vehicles is still running on Bharat Stage III over the study period.The emission factor for different categories of vehicles for this study is taken from CPCB (2010).
(a), the highest CO emission intensities for the year 2010 (more than 2000 kt) can be seen at different locations in Delhi viz.ISBT Kashmere Gate (3357 tons), ITO junction (3247 tons), I.P. Power plant (2945 tons), Karampura (2749 tons) and Shahazadabagh (2688 tons), which are due to relatively higher vehicular momentum, power-plants emission and industrial emissions.The minimum CO emission is observed at Kapas-Hera Estate (5.36 tons) in the year 2010.ISBT Kashmere Gate and ITO junction are the main hotspot traffic point of view in Delhi while Karampura and Shahazadabagh are the industrial areas.Emission from vehicular, industries, power plants and domestic sources of NO x pollution is shown in Fig. 2(b).

Fig. 2
Fig. 2(b).Spatial distribution of NO x emissions over Delhi urban area in 2010 (t/yr).

Fig. 3 .
Fig. 3. Comparison of CO, NO x and PM emissions of present study with previous studies.

Fig. 4 .
Fig. 4. Daily variations of observed concentrations for CO, NO x and PM at ITO junction.

Fig. 5
Fig. 5(b).Q-Q plots of observed and simulated NO x concentration.
as follows: Coefficient of Correlation (R): Coefficient of correlation (R) is a relative measure of the association between the observed and predicted values.It can vary from 0 (which indicates no correlation) to +1.0 (which indicates perfect correlation).A value of R close to 1.0 implies the total agreement between the observed and predicted values, i.e. best model performance.Error (NMSE): NMSE is the measure of performance, emphasizes the scatter in the entire data set and is defined as follows: It is a performance measure known as the normalized or the fractional bias of the mean concentrations: It is defined as: FAC2 = fraction of data that satisfy 0.5 ≤ C P /C O ≤ 2.0 where C P are the model predictions, C O are the observations, the overbar C is the average over the dataset, and σ C is the standard deviation over the data set.

Table 1 .
Correlation matrix of daily air pollutant concentrations and meteorological variables in Delhi for the year 2010.Temp-Temperature, DPT-dew point temperature, Hmd-humidity, Pres-pressure, Vis-visibility, WS-wind speed.