Characteristics of Surface Ozone Levels at Climatologically and Topographically Distinct Metropolitan Cities in India

Surface ozone (O 3 ) data at Pune (1998 - 2014) and Delhi (1998 - 2013) are studied to examine their temporal characteristics. Study also examines role of meteorology and atmospheric boundary layer height (ABLH) in modulating surface O 3 at these sites. Using diurnal variability of surface O 3 , rate of change of surface O 3 , [d(O 3 )/dt] is estimated to infer the nature of surface O 3 formation/destruction mechanisms. Analysis of data reveals that at both locations, surface O 3 concentrations during daytime are significantly high as compared to those during nighttime. Seasonally, at Pune averaged day-time surface O 3 concentrations are high during pre-monsoon and low in monsoon while those during winter and post-monsoon are found to be significantly higher than those in monsoon but half as compared to those in pre-monsoon. At Delhi, averaged daytime surface O 3 concentration is minimum in winter and maximum in pre-monsoon with monsoon and post-monsoon values being about 0.79 - 0.82 times with respect to pre-monsoon O 3 concentrations. High natural/anthropogenic pollutant concentration, abundance of ozone precursor gases, high temperature and high rate of photo-oxidation of precursor gases due to solar flux are the causal factors for increased surface O 3 concentrations in pre-monsoon season. Reduced solar flux decreases photo-dissociation of ozone precursor gases resulting in low O 3 concentration during winter season. Occurrence of low surface O 3 during early morning hours in monsoon, post-monsoon and winter seasons is because of low ABLH and low stratosphere-troposphere exchange (STE). [d(O 3 )/dt] values during morning/evening at Pune and Delhi are indicative of asymmetric and symmetric nature of ozone formation/destruction mechanisms.


INTRODUCTION
The stratospheric ozone layer can absorb a certain amount of ultraviolet radiation and protect the Earth's biosphere (Kutal et al., 2022;Lehman et al., 2004).However, near ground tropospheric ozone commonly referred as surface ozone, acts as an air pollutant to adversely impact human health and plant growth (Sharma and Sharma, reports (Resmi et al., 2020;Kalpana et al., 2019), claims that annually over 26 million Indian population receives treatment related to diseases arising from air pollution, making it 5 th chief reason for deaths in India.The geosphere and biosphere system at several parts of globe, especially Asia have witnessed numerous adverse environmental impacts due to increasing trend in surface ozone (O 3 ) along with its chief precursors methane (CH 4 ), non-methane hydrocarbons (NMHCs), nitrogen oxides (NOx), carbon monoxide (CO), along with other volatile organic compounds (VOCs) (Nair et al., 2018;Monks et al., 2015;Cooper et al., 2014;IPCC, 2013).Ozone being strong oxidizer reacts vigorously and heterogeneously with organic matter in biotic systems (Wang et al., 2015;Ghude et al., 2014;Lehman et al., 2004).The tropospheric ozone formation involves two distinct mechanisms: (i) transport from stratosphere to troposphere (Lelieveld and Dentener, 2000) and (ii) NOx catalyzed and controlled photochemical reaction via oxidation of VOCs, CH 4 , NMHCs and CO ( Jonson et al., 2006).Therefore, the ozone precursors and meteorological factors are important factors affecting ozone formation (Zhang et al., 2014).Investigations show that VOCs and NOx are the vital precursors (Wang et al., 2019;Zaveri et al., 2003).The rapid surge in industrialization, urbanization and anthropogenic activities is the chief reason for excessive emission of VOCs, NOx, CO and NMHCs into the atmosphere.These pollutants consequently with light exposure and other suitable meteorological conditions lead to surface ozone production with higher concentrations (Wang et al., 2014).Though local emission, ozone precursors largely determine tropospheric ozone concentration, since its lifespan extends to few weeks making it to get transmitted over long distances.Hence, at global level it is challenging task to control the surface ozone pollution.
Various researchers investigated the variations of surface ozone nearby coastal, oceanic regions and different tropical region from India (Girach et al., 2018;Peshin et al., 2017;Tyagi et al., 2016;Tiwari et al., 2015;Nair et al., 2011;Ghude et al., 2008).These studies largely specified that local emissions significantly influence the ozone and its precursors' seasonal distribution over the Indian subcontinent.However, in some cases there is disturbance in the variation of ozone that generally shows an afternoon peak.The secondary peak is observed during the night hours due to the enhanced ozone concentrations (Zhu et al., 2020;Wang et al., 2019;Kuang et al., 2011;Tong et al., 2011).This is common phenomenon all over globe (e.g., Europe, Asia) at all topographical sites.The ozone formation is ceased when solar radiation is absent, hence the likely reasons for enhancement of nocturnal surface ozone comes from vertical and horizontal transport phenomena (Zhu et al., 2020;Ghosh et al., 2013).The surface ozone concentration at any site could be the combined effect of the local meteorological conditions (viz., relative humidity (RH), temperature (T), wind speed (WS), and wind direction (WD) etc.) in association with dynamics of atmospheric boundary layer, its' stratospheric intrusion through stratospheretroposphere exchange, rate of photochemical production and long-range transport (Sharma, 2020;Sharma and Sharma, 2016;Nishanth et al., 2014).According to Gazette of India (Extra-ordinary Part-II Section 3, sub section (ii), dated Nov 18, 2009), the 1-hour and 8-hour average ambient air quality standards for surface ozone are 180 μg m -3 (~92 ppbv) and 100 μg m -3 (~51 ppbv) respectively (NAAQS, CPCB, 2009).The tropospheric ozone not only affect urban regions, but also, its transport from urban and industrialized areas to rural and remote oceanic regions plays vital role on human health.This type of transport (ozone and other pollutants) to Indian Ocean, Arabian Sea and Bay of Bengal from Indian subcontinent is revealed on the basis of ship-based measurements (Girach et al., 2017;Nair et al., 2011;Lawrence and Lelieveld, 2010;Lal et al., 2007Lal et al., , 2006)).
The rise in surface ozone concentration contributes to increase in global warming roughly by 10%, though the value is quite uncertain (Tropospheric Ozone Research (TOR) -2 final report, 2003; Saini et al., 2017).The ground and free tropospheric ozone concentrations at different locations all over the globe have shown both increasing and decreasing trends (Cooper et al., 2014;Kurokawa et al., 2009;Oltmans et al., 2008;Jaffe and Ray, 2007;Vingarzan, 2004).The yearly average surface ozone concentrations over Germany have doubled from 1950s till the end of 20 th century (Parrishet al., 2012).Similar trends have been noticed over European sites (Staehelin et al., 1994).A huge chunk of anthropogenic emissions of ozone precursors are being transported from Europe and North America to Asia since 1990s (Young et al., 2018;Zhang et al., 2016;Cooper et al., 2014;Granier et al., 2011).Recently, strict norms and control measures on ozone and its precursors have shown leveling off/ decrease in surface ozone in Europe and Eastern USA (Derwent et al., 2013;Oltmans et al., 2013;Parrish et al., 2012).The measurements show consistent rise in ozone and its precursors over Asian region (Nair et al., 2018;Xu et al., 2016;Akimoto et al., 2015;Munir et al., 2011).Recently Delhi, Capital of India, recorded very high surface ozone concentration exceeding the limits defined by World Health Organization (Anshika et al., 2021;Jain et al., 2005).The rise of 25-30 ppb in surface ozone in tropical India in this century is predicted by global chemical transport models (e.g.Brasseur and Solomon, 2006).The past few decades have witnessed rise in tropospheric ozone levels during spring over Northern Hemispheric mid-latitude remote sites (Parrish et al., 2013).Thus, it is imperative to analyze long term effects and trends of increasing surface ozone levels on environment and its bio-climatic significance.
The present study attempts to analyze the long-term trend of the quality-controlled surface ozone data at Pune (1998Pune ( -2014) ) and Delhi (1998Delhi ( -2013) ) received from India Meteorological Department (IMD).The key objective is to examine seasonal and diurnal scale variations of surface ozone.The data is further used to delineate the influence of the surface meteorological parameters on the behavioural pattern of surface ozone.The yearly data obtained from IMD has been segregated for the season wise analysis of surface ozone.

SITE DESCRIPTIONS, DATA AND METHODOLOGY
In this study, meteorological parameters (T, RH, WS) and surface ozone measurements received from IMD are analyzed over Pune (18.58°N,73.91°E,559 m AMSL) and Delhi (28.56°N, 77.11°E, 220 m AMSL) for the respective period of 1998 -2014 and 1998-2013.The heavy industrialization and urbanization makes these two cities obvious choice for the study since they record huge ozone precursor emissions (NOx and CO).These cities differ a lot climatologically as well in terrain, latitude, altitude, topography and surroundings (Fig. 1).
The Pune city is placed leeward sides of ranges of Western Ghat mountain with around 150 km stretched interior to the west of coast India.Pune city being the 8 th largest Indian metropolitan has witnessed a steep rise in urbanization mainly because of industrial growth (Pune city is information technology and automobile manufacturing hub of Maharashtra alongside it has many chemicals, foundries, food processing, construction engineering and electronic/electrical etc. industries) and rapidly increasing vehicular population in past two decades.Furthermore, being an educational hub (Oxford of the East) of India, the city witnesses immigration of large number students across the different parts of the country to peruse higher education.Pune is amongst one of the highly polluted Indian cities mainly due to the vehicular and industrial emissions.The climate of observational site, Pune, witness four seasons, signified with diverse air circulation/current patterns, according to the local geographical Indian system, namely monsoon season during June to September, post-monsoon for the period of October to November, winter all through December to February and pre-monsoon or summer during March to May.Postmonsoon is a transition season with a sporadic rainfall.There are seasonal directional changes in air masses which affect the observational area.
The megacity Delhi (the capital of India and one of the most polluted sites in the world) is situated at the central northern part of the Indo Gangetic Plains (IGP) (Anshika et al., 2021;Gupta, 2016).It has a huge vehicular traffic leading to enormous emissions of hydrocarbons, NOx, and CO.In its southwest direction the city has agricultural fields.Over Delhi, the temperature varies between very cold winters (average T = 12.9°C) and very hot summers (average T = 34.8°C),however, the average rainfall during monsoon season is around 825 mm (Sharma et al., 2016).Delhi experiences significant rise in surface ozone concentration with wide temporal and seasonal variation due to excessive precursor gaseous emission and subtropical atmosphere (Sharma et al., 2016;Ghude et al., 2008;Jain et al., 2005).
The study uses round the clock hourly measurements of surface ozone concentration and meteorological parameters (T, RH, WS, etc.) recorded by IMD at Pune and Delhi during study period.Surface ozone measurements were carried out with standard Potassium Iodide (KI) technique (Brewer electrochemical ozone sensor) (Chakrabarty and Peshin, 2016;Ali et al., 2012;IMD Manual, 1995).Sreedharan and Tiwari (1971) had reported the detailed operation procedure and instrument working principle.The accuracy of the surface ozone measurement by standard KI technique has been estimated to be ±10% (Chakrabarty and Peshin, 2016).The atmospheric boundary layer height (ABLH) data is obtained from the Modern-Era Retrospective analysis for Research and Applications (MERRA) model (https://giovanni.gsfc.nasa.gov/giovanni/) for Pune and Delhi.The surface ozone data with other meteorological measurements were analyzed statistically in this surface ozone study.

1. 1 Monthly Variation of Surface Ozone
Fig. 2 displays the variation of monthly averaged, daytime (08:00-20:00 hr)/nighttime (20:00-08:00 hr) averaged surface ozone concentrations (ppbv) along with the associated ABLH obtained from MERRA model at Pune (1998Pune ( -2014) ) and Delhi (1998Delhi ( -2013)).The vertical lines associated with each parameter in Fig. 2 correspond to ±1σ (standard deviation).Also, given in Table 1 are the values of maximum, minimum, mean and the median of surface ozone concentration (ppbv) at Pune and Delhi for the respective period of observations.Starting with the month of December, monthly mean surface ozone concentration at Pune increases from 7.94±5.18ppbv to attain its maximum value of 20.06±8.51ppbv in the month of April.From April onwards, surface ozone concentration depicts the decreasing trend to produce minimum values expect October for which it is 12.52± 6.44 ppbv (Table 1 and Fig. 2(a)).The daytime and nighttime surface ozone concentrations, delineating the differences in different months, reveal that the daytime averaged surface ozone concentration is found to be maximum in April (27.46±11.40ppbv) and minimum in Aug-ust (8.74±3.41ppbv).On the other hand, the nighttime averaged surface ozone concentrations appear to be maximum in April (12.85±6.67 ppbv)/May (12.72±6.40ppbv) and minimum in January (3.53±2.91 ppbv).From these data, it is clear that the maximum/minimum daytime surface ozone concentration is about 2.13/2.48times higher as compared to maximum/minimum nighttime surface ozone concentration.The daytime/nighttime averaged surface ozone concentrations reveal an interesting feature when the monthly data for these concentrations are grouped into conventional seasons.This procedure for Pune gives rise to higher daytime averaged surface ozone concentration during pre-monsoon season (26.65±9.72 ppbv) and lower concentration during monsoon season (11.26±5.32ppbv).The winter and postmonsoon surface ozone concentration values (13.89±  6.41 ppbv and 15.69±6.96ppbv respectively) are reasonably higher than their monsoon counterpart.Conversely, the nighttime averaged surface ozone concentration scenario over Pune is altogether different as compared to the daytime scenario, according to which, although, surface ozone concentration is found to be maximum in premonsoon, the monsoonal concentration appears to be ~1.5 and ~1.3 times lower than those observed during winter and post-monsoon.
The foregoing discussions reveal that the occurrence of relatively low surface ozone concentrations during monsoon as compared to winter/post-monsoon seasons and much lower as compared to those observed in pre-monsoon can be ascribed to the ozone deposition mechanism affected by humidity condition of underlying surfaces (Altimir et al., 2006).Further, high rainfall, low ambient temperature, high humidity and water vapour content in the atmosphere leads to the washout of the pollutants including ozone precursor gases (Sharma et al., 2013;Reddy et al., 2008).In addition to this, the overcast sky in monsoon season inhibits transmittance of solar radiation to the ground.In the presence of reduced sunlight in association with low surface temperature, the rate of the photochemical production of surface ozone gets subsided and hence less surface ozone production leading to low surface ozone concentrations during monsoon season (Ali et al., 2012;Debaje and Kakade, 2009;Beig et al., 2007).During pre-monsoon season, the pollutant concentration is at its peak due to anthropogenic and natural aerosol production activities in the atmosphere over Pune (Kolhe et al., 2016;Pawar et al., 2015).This may also include ozone precursor gases such as CH 4 , CO, NO/NOx, NMHCs, and VOCs (Trainer et al., 1987) which are abundantly present during pre-monsoon season in the atmosphere over Pune.The occurrence of high surface ozone concentrations during premonsoon season at Pune can be attributed to the prevailing high surface temperature (35°C-40°C) and the adequate presence of solar flux.This hastens the rate of photo-oxidation of precursor gases producing higher amount of surface ozone thereby leading to its higher concentration in the atmosphere over Pune.It is interesting to note that the surface ozone concentrations during winter and post-monsoon seasons are found to be significantly more than those in monsoon but practically half as compared to that in pre-monsoon.This predicts that the surface ozone variation is influenced both by the solar radiation and dynamical processes in the ABLH (Ahammed et al., 2006).During post-monsoon and winter seasons, after sunrise, ABLH gradually increases from 780-1,210 m to about 1,130-1,890 m around noontime due to convective activities (Fig. 2) with consequent reduction in the stratification of the ABLH.Around this time, the ambient air develops instability and rises due to surface heating produced by solar radiation.This accelerates the air pollutant dispersion and mixing of low ozone surface layer quantity with high ozone quantity in upper layer.This is the so-called Stratosphere-Troposphere exchange (STE) phenomenon which may be active during winter and post-monsoon seasons producing relatively higher surface ozone concentration in these seasons (Shan et al., 2008;Lal et al., 2000;Khemani et al., 1995;Levy et al., 1985).The effect of ABLH during monsoon and pre-monsoon seasons is also discernible in Fig. 2, according to which ABLH ranges between 630-880 m during monsoon season while it is found to be in the range 1,000-2,000 m during pre-monsoon season.As is evident from Fig. 2, the presence of low ABLH in monsoon is associated with low surface ozone concentration and high surface ozone concentration is linked with high ABLH during pre-monsoon.The collocated measurements of surface ozone and ABLH evolution over a tropical rural observing site, Gadanki (Andhra Pradesh, India) have established that the days of higher surface ozone concentration are associated with high ABLH and vice versa (Reddy et al., 2012).The present measurements corroborate with this finding.
As is seen from Table 1 and Fig. 2(b), at Delhi again, monthly mean surface ozone concentration is found to be minimum in January (7.92±5.52 ppbv) and maximum during April (15.87±11.34ppbv), May (14.75± 9.97 ppbv) and June (15.00±8.44 ppbv) months.In conformity with monthly means, the daytime minimum/maximum in averaged surface ozone concentration also occur in January (11.63±9.51ppbv) and April (25.09±12.11ppbv) months.The nighttime averaged surface ozone concentration also follows a little different trend in that it is minimum in November (3.50±3.29 ppbv) and maximum in June (7.66±6.13ppbv).Seasonally, the daytime averaged surface ozone concentration is minimum in winter (13.40±9.87ppbv) and maximum in pre-monsoon (22.86±10.89ppbv) with monsoon and post-monsoon values of averaged surface ozone concentration being about 0.79-0.82times with respect to pre-monsoon averaged surface ozone concentration.Earlier studies (Chakraborty and Peshin, 2016;Sharma et al., 2016;Ghude et al., 2008;Jain et al., 2005) have reported adequate build-up of surface ozone concentration giving rise to maximum surface ozone concentration during pre-monsoon season over Delhi which is highly industrialized and populated city situated in Indo-Gangetic plain of India.The plausible cause for significant increase in surface ozone concentration is the subtropical atmosphere of Delhi and large-scale emission of ozone forming precursor gases and the associated photo-chemistry.During winter most of the days are covered by intense fog which results in reduction of solar intensity to the greater extent (Safai et al., 2018).
The reduced intensity of solar radiation decreases the photo dissociation rate of ozone precursor gases resulting in low concentration of surface ozone during winter (Chakraborty and Peshin, 2016).In addition to the photochemical processes, the ABLH monthly variation at Delhi also plays pivotal role in controlling the monthly variability of surface ozone concentration as was discussed earlier for Pune.Thus, at both the sites, the photochemical surface ozone production mechanisms in the presence of sunlight and the dynamical processes in ABL have enormous implications on the monthly/seasonal variability of surface ozone which is true on diurnal scale also.In light of this, the diurnal variability of surface ozone concentration is discussed in subsections 3. 1. 2 and 3. 1. 3.

1. 2 General Features of Diurnal Variation
The physical and chemical mechanisms giving rise to diurnal variation are analyzed here so that the observational site can be monitored for different diurnal behavior characteristics.Fig. 3 depicts the contour plots of the averaged diurnal evolution of surface ozone for the study duration at the observing sites Pune (Fig. 3(a)) and Delhi (Fig. 3(b)) respectively.
The careful examination of Fig. 3 reveals that each contour diagram, depicting diurnal variability of surface ozone can be divided into three time zones based on the observed magnitude of surface ozone.The predominance of relatively higher surface ozone concentrations prevails during 09:00 to 17:00 hr with monthly/seasonally varying concentrations.At Pune (Fig. 3(a)), during winter season, the maximum ozone concentration is seen around 14:00 hr which lies in range 17.6-22.0ppbv while during pre-monsoon it appears to be higher in the range 26.7-30.1 ppbv.For monsoon season, however, concentrations almost become half (9.8-15.9ppbv) as compared to their pre-monsoon counter parts.There occurs a considerable increase in surface ozone concentrations during post-monsoon months which span in the range 19.6-22.9ppbv.The changes in prevailing air mass type, seasonal differences in meteorology and synoptic wind patterns attributes to seasonal variations (Gopal et al., 2014).In the morning time slot (i.e.midnight to morning 08:00 hr), in general, surface ozone concentrations are found to be low (less than 8 ppbv) for all months expect during April-June, for which it remains in the range 10-12 ppbv.Finally, in the evening time slot (i.e.beyond 19:00 hr to till midnight), once again ozone concentration becomes low and remain around 11 ppbv expect for April-June period where it reaches 24 ppbv concentration values.Similarly, at Delhi (Fig. 3(b)), for winter season, the maximum surface ozone concentration is seen around 14:00 hr which lies in range 18.8-22.7 ppbv while during pre-monsoon it appears to be higher in the range 28.9-33.1 ppbv.For monsoon season, however, concentrations depict decline in the range 20.8-27.9ppbv.There occurs a considerable increase in surface ozone concentrations during post-monsoon months which span in the range 28.7-34.1 ppbv.In the morning (i.e.midnight to morning 08:00 hr) and evening (i.e.beyond 19:00 hr to till midnight) time slots, surface O 3 concentrations remain less than 10 ppbv for all months.
The one-to-one comparison of surface ozone observed around 14:00 hr reveals that surface ozone concentration observed at Delhi is found to be comparable to that obs-erved at Pune during winter and pre-monsoon seasons.However, there exist a substantial difference between surface ozone concentrations measured at Delhi and Pune during monsoon and post-monsoon months.The surface ozone concentration at Delhi is about 1.7 ( June)-2.2 (July) time higher as compared to that observed at Pune.During post-monsoon months surface ozone concentration is about 1.5 times higher as compared to that observed at Pune.The observed difference between diurnal variability at Pune and Delhi can be mainly accounted to diverse regional/climatic zones represented by these two sites.Another reason for this variability could be the contribution of photo-chemically generated surface ozone from sunshine, different meteorological conditions, anthropogenic and natural precursors ( Jo and Park, 2005).The results presented in Fig. 3 for the diurnal variability of surface ozone at Pune and Delhi represents salient features, however this is an average picture.Therefore, in order to investigate more detailed features of the diurnal variability, it is essential to look into the causal factors responsible for producing it.

1. 3 Month-and Site-specific Diurnal Variations of
Surface Ozone Figs. 4 and 5 show the patterns of the monthly mean diurnal variability of surface ozone at Pune (1998Pune ( -2014) ) and Delhi (1998Delhi ( -2013) ) respectively.At the present observing sites, generally in all the months, the surface ozone is observed to be low in early morning which gradually increases after sunrise producing a broad maximum around noontime (Figs. 4 and 5).The occurrence of low surface ozone in early morning in all months/seasons is because of the low ABLH during early morning hours (especially in monsoon, post-monsoon and winter months) which reduces stratosphere-troposphere exchange (STE) (Reddy et al., 2008).A plausible reason for the high concentration during the afternoon hours is the photochemical production of excess surface ozone mainly from photo-oxidation of carbon monoxide (CO), industrial/anthropogenic hydrocarbons and methane (CH 4 ) in presence of enough amount of NOx and hydroxyl radical (OH) (Saini et al., 2017;Reddy et al., 2010;Seinfeld and Pandis, 1998).Thereafter, a gradual decrease in surface ozone is observed during nighttime and early morning hours.
The low surface ozone concentration during nighttime can be attributed to the absence of photolysis of NO 2 and the continuous loss of ozone (O 3 ) by NOx titration   is low during October-January while it starts increasing from February till May.During monsoon months (June-August) it starts subsiding with tendency to show its appearance again in September.Magnitude wise (Table 1), however, broad noontime maximum in the surface ozone concentration occurs in the range 9.79 ppbv (July)-30.16ppbv (April) while it is minimum (early morning around sunrise) in the range 2.92 ppbv ( January)-8.95 ppbv (May).The monthly diurnal variability at Delhi (Fig. 5) is found to be significantly different as compared to that observed at Pune.From Fig. 5, it can be noticed that surface ozone concentration during morning is comparatively less than that during nighttime.Further, it is noted to depict very small variation (of the order of 0.34 ppbv-2.68ppbv) during midnight to morning hours (up to 07:00 hr) while in the evening, from 19:00 hr to midnight, it shows significantly high variation from 1.47 ppbv-6.31ppbv.As against these variations, the noontime peaks are found to be prominent with sudden rise/ decay around 08:00/19:00 hr respectively.The major difference between the occurrence of broad peak at Delhi and Pune is that at Delhi the noontime broad peaks are well defined during all the months of year.

2 Rate of Change in Surface Ozone
The time rate of change of surface ozone [d(O 3 )/dt] defines the formation/destruction mechanisms of ozone.It can serve as a tool to distinguish between urban and rural observation sites on the basis of prevailing chemical environment over the site (Ojha et al., 2012;Naja and Lal, 2002).In general, for urban site, it is observed that the magnitude of [d(O 3 )/dt] during morning and evening hours is nearly same producing symmetric diurnal distribution of the surface ozone.On the other hand, at rural sites the [d(O 3 )/dt] values are less in evening hours than those in morning hours yielding asymmetrical diurnal variation (Ojha et al., 2012).In the present work, the magnitude of [d(O 3 )/dt] for Pune and Delhi are determined by employing the linear regression technique to the diurnal variation plots (Figs. 4 and 5) of surface ozone in morning (08:00-11:00 hr) and evening (17:00-19:00 hr) hours and the results are given in Table 2. From Table 2, it is seen that for both Pune and Delhi,[d(O 3 )/dt] values are positive/negative during morning/evening hours respectively corresponding to the formation/destruction of surface ozone.Table 2, further, reveals that at Pune, the averaged maximum/minimum [d(O 3 )/dt] values occur during March (6.06 ppbv hr -1 )/July (0.61 ppbv hr -1 ) while for Delhi the averaged maximum/minimum [d(O 3 )/dt] values are found to be during October (8.01 ppbv hr -1 )/January (2.54 ppbv hr -1 ) for morning hours.Alternately, for Pune during evening hours, the averaged maximum/minimum [d(O 3 )/dt] values occur during January (-4.20 ppbv hr -1 )/July (-0.58 ppbv hr -1 ) whereas at Delhi the averaged maximum/minimum [d(O 3 )/ dt] values are seen to be during April (-7.57ppbv hr -1 )/ November (-1.73 ppbv hr -1 ) respectively.

3 Comparison of Surface Ozone with
Other Sites Table 4 shows comparison of monthly average minimum and maximum levels of surface ozone concentration (ppbv) for 17 Indian sites (Verma et al., 2018;Shukla et al., 2017;Yadav et al., 2016;Malik et al., 2015;Gopal et al., 2014;Renuka et al., 2014;Sarangi et al., 2014;Sharma et al., 2013;David and Nair, 2011;Purkait et al., 2009;Lal, 2007;Debaje and Kakade, 2006;Debaje et al., 2003;Naja et al., 2003;Nair et al., 2002;Lal et al., 2000), 4 Chinese sites (Wang et al., 2012;Xu et al., 2008;Lin et al., 2008;Tu et al., 2007) and one site from Japan (Sikder et al., 2011).The monthly average minimum surface ozone concentration lies in the range of 2.83 ppbv-30 ppbv while maximum surface ozone concentration occurs between 22 ppbv-65 ppbv over Indian subcontinent.However, the monthly average minimum and maximum surface ozone concentrations over Chinese and Japanese sites lies in a range of 2.8 ppbv-23 ppbv and 33 ppbv-60 ppbv respectively.The differences in maximum and minimum surface ozone concentrations at different observing sites in India can be attributed to the difference in concentrations of precursor gases and their emission sources, chemical processes, natural and anthropogenic activities prevailing at the concerned site areas, the meteorology of the site under discussion and the influx of surface ozone via long-range transport at the respective sites.
The Pune and Delhi recorded lowest monthly mini-  (Purkait et al., 2009) across the India.At high altitude sites, the wind flow and ABLH affects the variation of both average minimum and maximum surface ozone concentrations rather than photochemistry of surface ozone (Sarangi et al., 2014;Sharma et al., 2013;Lal, 2007;Naja et al., 2003).At costal sites the broad daytime peak of ozone can be attributed to sea and land breeze (David and Nair, 2011).

4
Frequency Distribution of Surface Ozone Concentration Fig. 6 illustrates the percentage frequency distribution of surface ozone concentrations at Pune (1998Pune ( -2014) ) and Delhi (1998Delhi ( -2013) ) over all four seasons.There are total of 5,085 recorded data points for Pune site and 4,917 data points for Delhi site during the measurement period.However, the total number of observations made per season is shown in statistics so that one can estimate the actual frequency of occurrence of a particular concentration level.The different climatic conditions at Pune and Delhi may be the main cause of distinct seasonal frequency distributions of the surface ozone concentration.Based on observations and frequency distributions, it is evident from Fig. 6 that at Pune and Delhi, the maximum frequency of surface ozone occurrence during winter (47 and 54%) and post-monsoon (35.4 and 4.6%) seasons in the concentration range of 0-7 ppbv.Furthermore, at Pune, surface ozone frequency occurrence in the range 7-20 ppbv are found to be maximum during pre-monsoon (61.1%) and monsoon (61.8%) while, at Delhi, it is maximum during pre-monsoon (50.4%) and post-monsoon (36.3%) seasons.For higher surface ozone concentration (>20 ppbv), the maximum occurrence frequency is observed during premonsoon (30.5%) and post-monsoon (13.0%) seasons at Pune, while, during pre-monsoon (25.2%) and mon- soon (23.1%) at Delhi.The high surface ozone concentration, during the pre-monsoon season, is due to direct linear relationship with solar radiation exposure (temperature) and associated low relative humidity however, during the post-monsoon season blowing winds may be results in surface ozone dispersion (Kanchana et al., 2020;Verma et al., 2018;Singla et al., 2012).

5 Influence of Meteorological Parameters on Surface Ozone Concentration
The prevailing meteorological conditions at the place of observation site play dominant role on the surface ozone formation, its horizontal and vertical dispersion and transfer on short-and long-spatio-temporal scale.In the present work, the hourly average values of meteorological parameters T, RH, WS and surface ozone concentrations for Delhi and Pune are determined for the respective periods 1998-2013 and 1998-2014.The resulting average data is plotted in Fig. 7 on a seasonal basis as a function of hour which reveals that the param-eter T and RH are anti-correlated as expected; so also, RH is anti-correlated to surface ozone.Both T and RH are connected to precipitable water vapour content (PWC) in the atmosphere through the Leckner's formula as given in Eq. 1.
The Fig. 7 reveals a well defined seasonal variation of T, RH, WS and surface ozone on the diurnal scale.During all the seasons, the surface ozone follows similar diurnal variation pattern, however, the curve amplitudes and width of the presented curve are seen to be different.The statistics of the data represented in Fig. 7 is summarized in Table 5.At Pune, the highest mean concentration is during pre-monsoon (17.98±7.72 ppbv) followed by post-monsoon (10.87±6.15ppbv), winter (9.56±5.75At both the study locations, the high surface ozone concentration level observed during pre-monsoon season is attributed to the low RH, high T and long hours of sunlight (i.e.availability of intense solar radiation) which helps photochemical production of surface ozone (Kanchana et al., 2020;Verma et al., 2018;Yadav et al., 2016;Ali et al., 2012;Reddy et al., 2012;Wang et al., 2012;Tu et al., 2007).The presence of high RH/PWC in ambient atmosphere over the observing locations can produce depletion of surface ozone involving OH radicals for which PWC is the source (Nair et al., 2018;Gopal et al., 2014;David and Nair, 2011).In the present case, at Delhi, as seen above during monsoon season (Table 5), the surface ozone concentration is found to be low (0.84 times) as compared to that observed during pre-monsoon season and relatively high in post-monsoon (1.10 times) and winter (1.33 times) seasons.Here, these higher values during monsoon season may be partly attributed to the higher temperature (31.03±2.14)values over Delhi for this period in addition to other factors like local anthropogenic activities (vehicular exhausts, industrial/refin-  ery/power plant emissions, biomass/biofuels burning, etc.), ambient meteorology and processes in ABL (Reddy et al., 2012(Reddy et al., , 2008)).On the other hand, the net role of WS on the surface ozone depends on location and time.For example, at Delhi, for WS of 8.64±1.47km hr -1 , surface ozone concentration of 14.45±10.49ppbv is observed during pre-monsoon while at a reduced WS of 6.68± 1.82 km hr -1 , surface ozone concentration of 9.36±2.42ppbv is recorded during monsoon at Pune.

CONCLUSION
The present study analyses the temporal characteristics of surface O 3 on short-term basis using the datasets obtained from the National Data Center (NDC), India Meteorological Department (IMD), Pune during 1998Pune during -2014Pune during and 1998Pune during -2013 respectively over Pune and Delhi, known to be environmentally distinct metropolitan cities.The study also examines the influence of meteorological parameters (viz., T, RH/PWC, and WS) and ABLH data (retrieved from MEERA model) causing surface O 3 variability at the studied locations.The salient features of the study are as follows: • At Pune and Delhi, the averaged daytime surface ozone concentrations are found to be high during pre-monsoon while low concentrations occur during monsoon (Pune) and winter (Delhi).At both these locations, the maximum/minimum daytime surface ozone concentrations are found to be significantly high as compared to maximum/minimum nighttime surface ozone concentration.
• During pre-monsoon season, over Pune/Delhi, the concentrations of ozone precursor gases (CH 4 , CO, NO/NOx, NMHCs, and VOCs) are found to be at their peak due to anthropogenic and natural aerosol production activities in the atmosphere.Also, the adequate presence of solar flux and prevailing high surface temperature (35°C-40°C) hasten the photo-oxidation rate of precursor gases producing enhanced surface ozone concentration over Pune.However, the reduced intensity of solar radiation decreases the rate of photodissociation of ozone precursor gases resulting in low surface ozone concentration in winter.
• There exists a substantial difference between surface ozone concentration measured at Delhi and Pune during monsoon and post-monsoon months.The observed difference between diurnal variability at Pune and Delhi can be mainly attributed to the diverse regional/ climatic zones represented by these two sites, the contribution of photo-chemically generated surface ozone from sunshine, different meteorological conditions, anthropogenic and natural precursors as well as the influx of surface ozone via long-range transport at the respective sites.
• The occurrence of high RH (hence PWC) in the morning and evening hours is associated with low surface ozone concentration during morning and evening period.The presence of high RH/PWC in ambient atmosphere over observing locations can cause depletion of surface ozone involving OH radicals for which PWC is the source.This hypothesis supports the occurrence of low surface ozone concentration at high RH (morning and evening) and high surface ozone concentration (around noontime) at low RH over Delhi and Pune for the period of study exhibiting negative correlation between RH and surface ozone.Further, surface ozone concentration is found to be maximum for daily maximum temperature.
• The study demonstrates that the ABLH variation at Pune and Delhi plays a pivotal role in controlling the monthly/seasonal as well as the diurnal variability of surface ozone concentration at these sites.During postmonsoon and winter seasons, after sunrise, ABLH gradually increases around noontime due to convective activities with consequent reduction in the stratification of ABL.Around this time; the ambient air develops instability and rises due to surface heating produced by solar radiation.This accelerates the air pollutant dispersion and mixing of low ozone surface layer quantity with high ozone quantity in the upper layer.This process which is active during winter and postmonsoon seasons produces relatively higher surface ozone concentration in these seasons.

Fig. 3 .
Fig. 3. Contour plots of surface ozone concentration (ppbv) on a diurnal scale over the observation sites, Pune (Top panel) and Delhi (Bottom panel).

(
Debaje and Kakade, 2009).Also, the continual O 3 loss by the dry and wet deposition (the sink mechanism) produces minimum surface ozone in the early morning around sunrise.Further, the ozone production itself might be very low during nighttime/morning hours due to low precursor concentrations and reduced photochemical production mechanisms.Figs. 4 and 5 further depict varying amplitude of diurnal surface ozone distribution on month-to-month basis.At Pune, the monthly diurnal variability in surface ozone concentration depicts changing patterns from January to December months.However, the observed diurnal patterns during October-May months reveal somewhat similarity, although the duration i.e. width of the broad maximum around noontime significantly differs for these months.From Fig.4, it appears that the width of the broad maximum around noontime
averaged [d(O 3 )/dt] values for twelve environmentally different observing sites including Pune and Delhi sites which are explored in the present study.The annual average [d(O 3 )/dt] value during morning hours for Pune site is 3.6 ppbv hr -1 which is quite less compared to Delhi site (4.9 ppbv hr -1 ).Moreover, the [d(O 3 )/dt] averaged value
and lowest during monsoon (9.36±2.42ppbv) season.On the other hand, at Delhi, the mean surface ozone is found to be high during pre-monsoon (14.45± 10.49 ppbv) season and decreases with monsoon (12.11 ±7.46 ppbv), post-monsoon (11.04±10.41ppbv), and lowest during winter (9.1±6.3 ppbv) season.The occurrence of high RH (hence PWC) in the monsoon (80.68 % for Pune) and winter (72.06% for Delhi) is associated with low surface ozone concentration and vice-a-versa.

Fig. 7 .
Fig. 7. Averaged hourly variation of surface ozone concentration as a function of surface temperature (T, °C), percentage relative humidity (RH) and wind speed (WS, km hr -1 ) in winter, pre-monsoon, monsoon and post-monsoon seasons at Pune (a-d) and Delhi (e-h).

Table 1 .
Detailed statistics of surface ozone (ppbv) including monthly averaged surface ozone concentrations with ±1σ standard deviation, maximum, minimum, and median in each month for the study period at Pune and Delhi.

Table 2 .
Month-wise rate of change of surface ozone (ppbv/hr) during morning and evening hours.

Table 3 .
Comparison of rate of change of ozone (ppbv/hr) at study locations with other sites in India.

Table 4 .
Comparison of monthly average minimum-maximum levels of surface ozone (ppbv) with other sites.

Table 5 .
Seasonal statistics of surface ozone, relative humidity (RH), temperature (T), and wind speed (WS) observational data shown in Fig.7at study locationsPune and Delhi.