Enhanced Ozone Production in Ambient Air at Patiala Semi-Urban Site During Crop Residue Burning Events

The continuous in situ measurements of criteria pollutants (O3, CO and NOx) and meteorological variables were done at a semi-urban site (Patiala) of North-West Indo-Gangetic Plains (NW-IGP) for the years 2014 and 2015. The influence of meteorology on the levels of criteria pollutants was investigated on daily, monthly, seasonal and diurnal basis. The average rate of increase in O3, calculated to be 1.1 ppb h−1 and pollution index (daily O3 max/O3 min ratio), values (10.2) indicates remarkable ozone pollution. Ozone is negatively correlated (represented as r value) with its precursor gases CO (− 0.28) and NOx (− 0.32), as they get consumed in the photochemical production of ozone. The impact of meteorology on ozone production was positively correlated with SR (0.63), AT (0.49), MT (0.59) and WS (0.23) and negatively correlated with RH (− 0.83). Pre-, during and post-biomass burning periods were determined using Terra-MODIS images over the study area. Enhanced levels of ozone were recorded as 20 ppb and 15 ppb during daytime, respectively, for rice and wheat crop seasons, and 12 ppb during nighttime for both seasons. Ozone exceedance of 24 h national standard occurred on 50% and 8% of the sampling periods during wheat and rice crop residue burning, respectively. Ozone generation with its precursor gases was analyzed quantitatively.


Introduction
Agriculture is the backbone of the Indian economy with an average growth rate of approximately 7% over the last two decades [1]. Indo-Gangetic Plain (IGP) is one the world's largest and agriculturally productive areas ranging from 21.75°N, 74.25°E to 31.0°N, 91.50°E. It accounts for 21% of land area of the Indian subcontinent and holds nearly 40% of the total population [2]. This densely populated area has experienced extensive economic growth, urban expansion and development in various sectors, i.e., industry, automobile, infrastructure, agriculture, etc. which resulted in pollution problems in the past few decades. Natural and anthropogenic activities emit aerosols and trace gases such as methane (CH 4 ), total non-methane (TNMHCs), carbon monoxide (CO) and oxide of nitrogen (NO x ), which impart significant role in the photochemical production of surface ozone, involve oxidation of its precursors (volatile organic compounds (VOCs) and CO), are catalyzed by OH radical and are controlled by NO x . The detailed photochemical reactions of the production of surface O 3 have been provided by [3,4]. The rate of increase in tropospheric ozone per decade in northern India is more than twice that in the southern India, and residence timescale also varied with altitude and season to season [5]. Tropospheric ozone, a greenhouse gas, a powerful oxidant contributes radiative forcing in atmosphere and also imposes a big threat for crop productivity and respiratory problems of the human beings [6][7][8].
The studies of the tropospheric chemistry of geographical areas (Africa, South America, and South Asia) are emphasized due to the global impact of biomass burning on air quality and climate change. Extensive agricultural crop residue burning (CRB) occurred in the IGP region-the most intensely farmed zones in the world. Rice and wheat are the major crops, which generate a large amount of leftover crop residue, intensified with the usage of mechanical combined harvester technology which is further subjected to open field crop residue burning [9,10]. The harvesting of crops is done in the months of April and October for the crops of wheat and rice, respectively, whereas the burning of crop residue from harvesting is done from mid-April till May and mid-October till mid-November for the respective crop seasons. The post-harvest wheat crop residue burning (WCRB) during April-May and rice crop residue burning (RCRB) during October-November are the conspicuous features in the N.W IGP. Burning of the crop residue in the open fields becomes the prime source of the micron-sized aerosols [11,12] and trace gases [10,13], which due to long-range transport affect the composition and reactivity of atmosphere and more deep inversions [14] and health of individuals [15].
Few studies were reported on the influence of agricultural crop residue burning (CRB) on surface ozone levels over IGP region [16][17][18][19] which leads to critical importance for continuous, long duration in situ monitoring of trace gases. Under the Government of India-funded project, Modelling Atmospheric Pollution and Networking (MAPAN) initiated by IITM, Pune, continuous in situ measurements of ozone and its precursor gases are monitored at Patiala, a semi-urban site in India. The aim of the present study is to probe the impact of emissions and meteorology on the diurnal and seasonal variability of O 3 and its precursor gases (CO and NO x ) with an emphasis on regional agricultural crop residue burning on ambient air quality. Agricultural crop residue burning in the fields is a widespread practice and was estimated along with their impact on ambient air quality during non-burning and burning periods. A high-resolution, quality-controlled 2-year-long in situ data set of ozone, nitrogen oxides and carbon monoxide was monitored from January 2014 to December 2015 at a regionally representative semi-urban site in the N.W. IGP. The station is installed on the top of the second floor (50 feet height) from the ground level, and the inlet system of analyzers and samplers is installed on the rooftop. The geographical location of monitoring site is shown in Fig. 1. University campus has a sequence of buildings (academic and residential houses) concatenated by roads with trees planted on both sides and scattered grassy lawns and vehicle parking areas. As per data of 2011 census, Patiala City had a population of 4,06,192. Patiala is situated in the northwest part of the Indo-Gangetic plain, close to Shivalik Hills in the east and the Thar Desert in the southwest. It is circumscribed by district Ambala . The climate of the city is very hot in the summer (with a maximum temperature of 43°C during May) and very cold in the winter (minimum temperature of 2°C during January). The monsoon season contributed to the annual rainfall of 600 mm. This region is also affected by the prominent synoptic meteorological phenomenon (western disturbances) that is dominant during pre-monsoon and winter seasons.

Study Area and
Rice and wheat are the two major cereal crops of this region with a combined cropping area of more than 78%, which is 12% of the cereals produced in India. The largest grown crop is wheat followed by rice, and other important crops are cotton, sugarcane, pearl millet, maize, barley and fruits. Among different crops, cereals generate a lot of crop residues, followed by fibers, oilseeds, pulses and sugarcane [9]. Agricultural crop residue generation rises sharply with the rise in production after the adoption of green revolution technologies, high-yielding varieties, use of chemical fertilizers, irrigation facilities, new machinery and expansion of crop areas. The total crop residue generated in India is highest in Uttar Pradesh followed by Punjab and Maharashtra.
In the northwestern part of the IGP (Punjab, Haryana and Western Uttar Pradesh), wheat and rice are harvested in the months of April and October, respectively. This activity leaves behind a lot of crop residue which is burnt to clear the fields for the next crops. The farmers adopt the easiest way, i.e., burning of the crop residues in the open field for crop rotation due to short duration between rice harvesting and wheat sowing, labor shortage, use of mechanized combined harvesting technology, uncertainty of weather, and no economical technologies are used for collecting leftover agricultural residues from the field [12]. The burning of wheat and rice crop residue is done from mid-April till May and mid-October till mid-November for respective crop seasons. Open field crop residue burning due to less than ideal combustion condition typically produces smoke which contains harmful gases and pollutants during burning episodes that ultimately affect the human health.
Pre-, during and the post-biomass burning period for wheat and rice has been outlined over the study period using Terra-MODIS (Moderate Resolution Imaging Spectroradiometer) true color composites provided in supporting material ( Fig. SM-1). Terra-MODIS (Moderate Resolution Imaging Spectroradiometer) satellite observations [20] clearly indicate a large number of fire spots (red dots) during wheat and rice crop residue burning period and related emissions in the form of smoke over the IGP region. For better understanding and comparison of wheat and rice crop residue burning, the study period was identified with the support of MODIS true color composites as: wheat crop residue burning (WCRB), designated as; prewheat crop residue burning (15th March to 31st March), during wheat crop residue burning (20th April to 31st May) and post-wheat crop residue burning (15th June to 30th June). Rice crop residue burning (RCRB) is designated as pre-rice crop residue burning (1st September to 15th September), during rice crop residue burning (15th October to 20th November) and post-rice crop residue burning (1st December to 15th December).

Instrumentation
Data analyzers used in the system are based on USEPA equivalent referenced methods and are connected digitally to Data Acquisition System (DAS) having attributes like automatic data validation software. In this study, daily air quality data of O3, CO and NO x were collected for the 2-year period of January 2014 to December 2015 and used for the study. The one hourly averaged data of surface meteorological variables [average temperature (AT), solar radiation (SR), relative humidity (RH), wind speed (WS) and wind direction (WD)] were retrieved for the given duration.
In the present study, surface O 3 is measured by an analyzer (Serinus 10, Ecotech), which works on the UV light absorption principle (UV photometry). It automatically rectifies changes in temperature and pressure, referenced to 0°C, 20°C or 25°C at 1 atmosphere within a range and precision of 0-20 ppm and 0.5 ppb, respectively. It determines the concentration of ozone in a sample gas at ambient pressure by detecting the absorption of UV radiation at 253.7 nm by O 3 molecule. The sample gas and the O 3 -free reference gas are passed through UV light absorption cell, and the intensity is measured by the light detector. The difference between the intensity of the signals is used in the basis for computing the O 3 concentration, based on Beer-Lambert Law. NO x measurement is based on gas phase chemiluminescence method using an (Serinus 40, Ecotech) analyzer with a precision of 0.4 ppb in the range of 0-20 ppm. Sample air is drawn into the reaction cell via two separate (alternating) channels the NO and NO x . The NO x channel travels through a delay coil enabling the same sample of air to be sampled for NO, NO 2 and NO x . The NO x channel passes through an NO 2 to NO converter, and NO 2 is converted to NO. Sample air (NO and NO x channels) enters the measurement cell where NO reacts with ozone in the following reaction This reaction releases energy in the form of chemiluminescence radiation (1100 nm), which is filtered by the optical bandpass filter and detected by the photomultiplier tube (PMT). The level of chemiluminescence detected is directly proportionally to the NO in sample. NO 2 is calculated by subtracting the NO measurement from NO x measurement.
CO measurement is based on the infrared (IR) photometry using the gas filter wheel correlation technique using an (Serinus 30, Ecotech) analyzer within a range and precision of 0-200 ppm and 0.05 ppm, respectively. CO absorbs infrared radiation (IR) at a wavelength near 4.7 microns. IR radiation (at 4.7 microns) is passed through a 5 m path length through sample air. The strength of the signal received is proportional to the amount of CO in the sample, based on Beer-Lambert Law. The daily zero/span checks were done using Ecotech Gas Cal 1100, a gas calibrator [21]. The maintenance and calibration protocol for O 3 , CO, NO x analyzer, meteorological sensors and gas calibration system are mentioned in detail in Table SM-1. The mass concentration was observed at every 5-min interval and then stored in the data repository as 1 h average. It performs an automatic check of its calibration (span check) and instrument drift caused by varying temperature, barometric pressure and relative humidity on an hourly basis during operation.

Results and discussion
The measurement of primary and secondary gaseous pollutants helps to study the changes in atmospheric chemical composition and to understand the ambient air quality. This study presents a 2-year (2014 and 2015) data of continuous measurement of ozone, CO, NO x and meteorological variables (average temperature, solar radiation, relative humidity, wind speed and wind direction). The data coverage for the entire study period was very high (about 95%) and hence statistically robust. Figure 2a shows the daily mean, maxima and minimal levels of gaseous pollutants (O 3 , CO and NO x ) and daily mean variation in meteorological variables (RH, SR, AT, T max and WS) for the study period at Patiala. The daily averaged mean gaseous pollutants O 3 , maximum O 3 , CO and NO x concentrations ranged from 4.9 to 80.9 ppb, 11.2 to 144.2 ppb, 0.05 to 2.19 ppm and 2.3 to 63.2 ppb, respectively, and meteorological variables RH, SR, AT, T max and WS varied from 33.6 to 95.0%, 69.2 to 377.7 W m -2 , 8 to 38°C, 9 to 46.8°C and 0.1 to 7.9 m s -1 , respectively. Moderate to high values of the gaseous pollutants were reported for other rural, semi-urban, urban, coastal and mountain sites in India [10,12,16,[23][24][25][26][27][28]. Figure 2b, c shows concentration levels of monthly average calculated from 24-h data set for gaseous pollutants (O 3 , CO and NO x ) and meteorological variables (RH, SR, AT) for 2014 and 2015. The increase/decrease in monthly concentration levels of each gas indicates the significant influence of human activities, seasonal variations and related chemical transformations. A maximum of ozone concentration level is corresponding to the minimum levels in precursor gases like CO and NO x that lead to the formation of ozone, with favorable meteorological conditions. The daily O 3 max /O 3 min ratio is a pollution index; for polluted sites, it has values in excess of 10 while low values smaller than 1.40 for clean areas [29]. The index value of the order of 10.20 indicates the study location has remarkable ozone pollution. The comparison of levels of ozone in the ambient air of Patiala site with that of some other locations in India is reported in Table 1. Figure 2b shows that ozone level starts increasing from January 2014 and the first maxima appeared in May (56.8 ppb) followed by decreases till September with a minimum value of 28.3 ppb in monsoon season. Second maxima appeared in November 2014 with a level of 41.8 ppb followed again by a decrease till January 2015.
The same trend of variations in the level of ozone was observed in 2015 with a little change in magnitude of the average concentration levels.
In the given study period, ozone levels showed maxima during each crop residue burning (CRB) period and corresponding minima's in precursor gases as they got consumed during the formation of secondary pollutant, ozone. High amplitude peak of ozone during wheat crop residue burning (WCRB) is the result of favorable meteorological conditions, i.e., high ambient temperature, solar radiation and low relative humidity. During rice crop residue burning (RCRB) period, burning proceeds at low temperature which leads to high emissions of precursor gases, but still a low amplitude peak was recorded as compared to WCRB period, which is due to low solar radiation intensity during the months of October and November, which is essential for photochemical synthesis of ozone. The effect of this episodic event is explained in detail in later section.  Fig. 3a. In diurnal ozone profile (Fig. 3a), a steep minimum is observed during the early morning around 0600-0800 h in all seasons of the year. After 0800 h, a rise in ozone concentration is observed and hits maxima around in the range of 1500-1600 h and starts decreasing after late afternoon or evening. However, significant ozone concentration is found throughout the study period with changes in its amplitude during different seasons. During daytime, photochemical reaction of precursor gases in the presence of favorable meteorological conditions and vertical mixing from ozone rich upper layer contribute to surface ozone levels [25]. During early morning and night hour, low boundary layer height reduces the vertical mixing process and the higher concentration of NO and NO 2 leads to rapid titration of ozone. The average O 3 values during daytime (0600-1700 h) and nighttime (1800-0500 h) for the study period (Table SM-2; in supplementary information) were used to estimate the average rate of increase of O 3 , calculated to be 1.09 ppb h -1 .
Ozone diurnal variation showed a similar pattern with varied amplitude, from season to season. A pronounced and sharper increment in daytime maximum ozone is observed during the post-monsoon season, whereas the daytime    precursor gases, especially from the crop residue burning event [19]. The average rates of increase/decrease [d(O3)/ dt] were as high ?8.5 ppb/h as compared to different cities of India, e.g., Delhi [23], Agra [18], Kanpur [16], Udaipur [28], which lies between (2.5 and 5.9) ppb/hr during 0800-1100 h and -8.1 ppb/h lies between (-2.4 and -6.4) during 1700-1900 h. The rate of change of ozone of surface ozone at any site is dependent on its net chemical production (production loss), surface deposition and advection. Figure 4a, b shows diurnal variations (seasonal average values) of NO x and CO during different seasons of the study period, and peaks were observed in early morning and late evening hours. On a diurnal scale, precursor gases exhibited two maxima, at around 0700-0900 h and 1800-2200 h, corresponding to low boundary layer and prominent anthropogenic emissions during evening hours, when the vehicular activity is maximum. The rise in boundary layer during 1400-1700 h provides favorable dispersion conditions coupled with reduced vehicular emissions as compared to the rush hours resulted in minimal levels of CO and NO x . Apart from the combustion sources, another source for CO is formaldehyde (HCHO) which is generated from the photochemical reaction of hydrocarbons. HCHO accumulates during night time and photo-dissociates into CO during daytime [30]. The chemical production of NO x (NO 2 ? NO) is mostly from the photo-dissociation of NO 3 and N 2 O 5 . In ozone production, NO 2 gets photo-dissociated and CO reacts with OH radical during sunshine hours, which leads to minimum levels of NO and CO [16].
Comparison of daily ozone profiles with those of precursor gases (NO x and CO) showed that maximum O 3 concentration was coincident with the concentration minima of NO   Figure 5 shows the dependence of O 3 on the levels of CO and NO x during daytime (0800-1700 h) for different seasons. During premonsoon/summer and post-monsoon seasons, the enhancements in levels of O 3 (as indicated by intensity of red color) were observed in the low-CO and low-NO x regimes, which is a result of the rise in the photochemical production of O 3 . The large production of ozone is also indicated by a large number of data points in this region, while in the winter season, the low levels of ozone are spread over the whole contour plot indicating continuous production of ozone as reflected from yellow to blue color. For low-NO x regime, increase in NO x leads to increase in OH and to corresponding increases in the oxidation rate of hydrocarbons and in the levels of ozone the rate is independent of CO. In the high-NO x limit (NO-saturated), the rate of ozone formation increases with CO but decreases as NO increases [31].
Box and Whisker plots in Fig. 6a-c show the seasonal averages of ozone, CO and NO x levels for the period of 2 years (2014 and 2015). The vertical lines stretched out from each end of the box called whiskers, representing 1% (bottom cross) and 99% (top cross) of the data. The vertical box encloses the minimum (bottom line), maximum (top line) and middle line 50% of data. The mean value represented by the cross-and horizontal line inside the box stands for the median value. There is very distinct and systematic seasonal variation observed in the levels of ozone, CO and NO x . Seasonally averaged ozone concentrations can be put in decreasing order as: summer [ post-monsoon [ monsoon [ winter. The observed changes in ozone concentrations are regulated by various interlinked processes of atmospheric chemistry, daily variations in boundary layer height, dynamics and transport of pollutants. Among all seasons, ozone production is intense in summer, because of increased intensity of solar radiations during the day as well as strong and persistent inversions at nights. In monsoon, low levels of ozone were observed because of wet surface deposition of precursor pollutants by monsoonal rains and limited solar radiation intensity. NO x and CO emission sources are anthropogenic in nature and followed similar seasonal trends in concentration levels in the order: winter [ post-monsoon [ summer [ monsoon. Minimum ozone is produced in winter due to the low solar intensity and low temperature while there is not much fall in anthropogenic emissions of precursor gases (CO and NO x ) in winter followed by the post-monsoon season. This illustrates that the photochemical production of ozone resulted from reaction of CO, CH 4 and NMHCs with OH radicals in the presence of NO x . The long-range transport, atmospheric reactivity, regional meteorology and emission sources play important role in variation in gaseous pollutants at Patiala. The 5-day backward trajectory at 500-m level over Patiala for four different seasons has been plotted in order to show the more realistic motion of air parcels and their transport pathways, as shown in Fig. 7. The substantial sources of emissions at this site are local vehicular exhaust and agricultural crop residue burning in the surrounding regions. In pre-monsoon/summer season, long-range transport of air masses from desert regions of Iran, Afghanistan, Pakistan and Arabian Peninsula contributed to aged pollutants and regional emissions of wheat crop residue burning (WCRB) event contributed to the increased emission levels of precursor gases (like NO x and CO) for ozone production, under favorable meteorological conditions. In monsoon season, trajectories indicate transport of cleaner air from the Indian Ocean and Arabian Sea due to SW winds. The levels of pollutants show a significant impact of washout due to dilution with the transport of cleaner air, frequent rainfalls, decreased solar radiation and negligible contribution from the agricultural crop residue burning events during monsoon compared to other seasons. In post-monsoon, trajectories show the mixed air masses (continental and marine) and the transition from SW (during monsoon) to NW majorly. High levels of emissions were reported due to regional rice crop residue burning (RCRB) event, festival period and the advent of low temperature and calm wind weather conditions. In the winter season, regional air masses reported high levels of CO and NO x and low levels of ozone, due to decreased solar radiation intensity and sunshine hours, which inhibits the photochemical processes. The low boundary layer and calm meteorological conditions further limit the dispersion of regional emissions of biomass burning and vehicular exhaust [32].

Dependence of O 3 on Precursors and Meteorology: Correlation Analysis
A parametric, Pearson correlation test is advantageous for the analysis of quantitative variables with different units. The complex relationship between O 3 and other variables that may induce or impede its production was analyzed using this method. Correlations between gaseous pollutants (O 3 , CO and NO x ) and meteorological variables were computed using SPSS Ò , IBM Ò Ver. 20 (Statistical Package for the Social Sciences). Significant correlations (p \ 0.01) were observed in all cases as shown in Table 2. Further, quantitative analysis of gaseous pollutants with significantly correlated meteorological variables was also done. Ozone exhibits strong day to day, seasonal and longterm timescale variations which are determined by anthropogenic processes and meteorological conditions since, CO and NO x are the precursor gases for the production of ozone under photochemical conditions. The rate of ozone production is negatively correlated with the levels of CO and NO x in the tropospheric region. The impact of meteorology on ozone production can be observed as it is positively correlated with SR, AT, MT and WS and negatively correlated with RH. The highest positive correlation coefficient of 0.628 was observed between SR and ozone which highlight the importance of photochemistry in the formation of ozone. A linear relationship between daily average temperature (AT), daily maximum temperature (MT) and ozone leads to higher ozone concentration, with heat accelerating the rate of atmospheric chemical reactions. The negative correlation (r = -0.826) between the surface ozone and humidity indicates that with an increase in humidity levels, the precursor gases are washed away from the photochemical paths resulting in fall in production of tropospheric ozone concentration.
Higher levels of humidity are associated with atmospheric instability, greater cloud abundance and reduced incoming solar radiation, which diminishes the photochemical processes. Another factor for surface ozone reduction is the deposition of water droplets. The positive correlation (r = 0.23) between ozone and wind speed points out to ozone transport. The increase in wind speed implies the increasing transport of air, ozone concentration increased due to its transport from other regions [33], and low wind speed promotes the buildup of high local ozone concentrations.
CO and NO x showed a strongest positive correlation of r = 0.748, attributable to common emission sources such as road traffic and combustion sources. CO and NO x were negatively correlated with SR (-0.532, -0.569), AT Table 2 Correlation coefficient (r) between gaseous pollutants (O 3 , CO and NO x ) and meteorological variables (relative humidity (RH), solar radiation (SR), average temperature (AT), maximum temperature (T max ) and wind speed (WS) for the study period

Quantitative analysis
The quantitative analysis emphasizes on the study of dependencies of gaseous pollutants concentration levels in a range of highly dependent meteorological variables. Table 3a, b summarizes average concentration levels of gaseous pollutants determined within temperature and relative humidity ranges, respectively. The relationship between temperature and gaseous pollutant can be easily explained on theoretical grounds. Temperature determines the rate of radical chain propagation which is important for photochemical ozone formation reaction [34].
Ozone concentration increased with increase in temperature, and the highest average levels of ozone occurred at an ambient temperature greater than 35°C. On the contrary, the O 3 precursors showed peaks within lowtemperature ranges that prescribe their convective dispersal and diminished atmospheric photochemical reactivity. As temperature increases, CO and NO x get consumed in the production of ozone. As shown in Table 3b, the highest average concentration of ozone occurred at humidity less than or equal to 40% and vice versa for CO and NO x due to enhanced oxidation of hydrocarbons in the afternoon, which is associated with the ozone production [35].

Case study: Ozone Levels During Crop Residue
Burning (CRB) Episodes at a Semi-Urban Site A preliminary analysis of the plots (Fig. 2a, b) indicates that ozone levels are a reflection of the trends of CO and NO x , besides being more susceptible to crop residue burning (CRB) activities. Hence, further discussion is confined to the ozone levels for the above said period. With a view to understanding variations in ozone levels, especially during the episodic CRB event, pre-, during and post-biomass burning period has been depicted over the study period using Terra-MODIS true color composites shown in Fig. SM-1. The clear increment in the level of ozone can be seen during the wheat and rice residue burning as shown in Fig. 8a-d.
The wind rose diagrams for the corresponding periods, that is, pre-, during and post-CRB events, have also been plotted (Fig. 9) and taken into consideration while analyzing the data. These wind plot diagrams represent the wind direction with a frequency of wind blowing from along with the wind speed based on hourly wind data. The frequency of the time for which wind blows from a particular direction is represented by the length of each ''spoke.'' Each concentric circle represents a different frequency, starting from zero at the center to increasing frequencies at the outer circles. It is observed that most prevalent wind direction at the site is from the NW in wheat and NNE and NE in rice period. Almost similar fetch regions were observed in pre-, during and post-harvesting periods of wheat and rice, respectively. Calm conditions were observed more often in during and postperiod of rice crop residue burning (RCRB), whereas high wind speeds (2-8 m/s) were observed in the wheat period. The pre-harvesting periods can be considered as ''baseline periods'' for evaluating the impacts of crop residue burning on ambient ozone concentration levels.
The levels of primary gaseous pollutants (CO and NO x ) are almost double in RCRB period than in the WCRB. The probable reasons are complete rice stubble as it is not used for animal feed while wheat stubble is stocked for animal feed; combustion process occurred at low intensity of solar radiation and calm wind conditions as compared to WCRB which influence the chemical reactions of pollutants and their dispersion [9,12,38]. In WCRB period, the primary pollutants (CO and NO x ) got consumed in photo-oxidation reaction for the production of secondary pollutant, ozone under favorable meteorological conditions. Figure 8 shows bar charts for ozone levels for pre-, during and post-CRB events. Although separate bars have been plotted for ozone levels recorded during the daytime, nighttime and 24 h average, trends in variations are more or less the same. The daytime data as seen from bar charts (the values during harvesting season) are significantly higher than the pre-and post-CRB period. Similar trends are noticed for ozone levels of RCRB-2014 and RCRB-2015 except for pre-RCRB, where relatively less concentration of ozone was observed in 2014 than 2015 due to rainy days during that period.
The level of ozone increased by 25% during WCRB period (2014 and 2015) and the maximum monthly average were observed in May (56.78 ppb in 2014 and 53.53 ppb in 2015). The peak levels of averaged daily O 3 increase from 49 ppb in pre-CRB to 72 ppb during CRB period in 2014 and from 41 ppb to 67 ppb in 2015. The available hour of solar radiation is very similar during both periods (pre and during); the noteworthy increase was driven by the difference in the intensity of solar radiation (increased by 50 W m -2 ) and enhanced emissions of ozone precursor gases from the crop residue burning. Ozone exceedance days were observed during WCRB (20 days) and post-WCRB (15 days). This episodic pollution event extends till post-CRB period due to prevailing calm atmospheric conditions. In case of rice, low levels of ozone were monitored during RCRB period as compared to wheat, with an increase of 10-20 ppb from pre-RCRB periods. Despite profound emissions of precursors, such as CO and NO x from CRB, the conditions for the photochemical formation of ozone are unfavorable. Figure 2c shows the consistent decrease in temperatures and solar radiation during this period. The increase in levels of ozone was observed, approaching the standards, but only 3-5 ozone exceedance days were observed on the 8 h-max basis and condition persist till post-RCRB due to calm atmospheric conditions and low boundary layer in the winter season. The CRB event influences the ozone levels in a sustained manner and influences positively for a period much beyond 1 month as the ozone production doesn't come back to the pre-CRB period levels. As seen from the wind rose diagram, the fetch region was distributed from SW to NW in WCRB and N to NE in RCRB. In conclusion, we can safely interpret that the remarkable increase in pollution levels was done due to episodic CRB event. The fetch regions were remained almost same throughout the study of pre-, during and post-CRB, neutralizing the long-range transport of pollutants.

Conclusions
The influence of meteorology on continuous temporal and spatial levels of ozone, CO and NO x was investigated at a semi-urban site of Patiala, India, during 2014-2015. Maximum O 3 concentrations were coincident with the concentration minima of NO x and CO on the diurnal scale. Seasonally averaged ozone concentrations showed an order of summer [ post-monsoon [ monsoon [ winter. As expected, ozone concentration levels showed an increase with an increase in daytime temperature with ozone maxima observed at 35°C or more. Complimentary minima for precursor gases (CO and NO x ) were observed during the temperature greater than 35°C. The highest average concentration of ozone recorded during noon time at humidity B 40% is attributed to intense oxidation of hydrocarbons at that time, which aid the ozone production.
The average rate of increase of O 3 was observed at 1.1 ppb h -1 , and during crop residue burning periods the almost double rate was observed. The National Ambient Air Quality Standards for ozone (60 ppb for 24 h) were violated 50% and 8% of the duration during wheat and rice crop residue burning, respectively. Seasonal backward trajectories highlighted the contribution of regional emissions as compared to long-range aged pollutants during crop residue burning period. Crop residue burning events influenced the ozone levels positively in a sustained manner for a period much beyond the post-harvesting season. Regulation to curb crop residue burning would surely extenuate ozone pollution and bring co-benefit in productivity of crops.