Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
 We study the effect of lightning activity on surface NOx and O3 over Pune, India (18.54°N, 73.81°E) for the years 2005 to 2008 during thunderstorm events in premonsoon and monsoon periods. Surface concentration of NOx is found to be increased significantly at the dissipation stage of thunderstorms. It is observed that increase in NOx greater than titration threshold level reduces the surface ozone concentration. However, in some cases when NOx increases but it does not reach the titration threshold limit, it helps in the production of ozone. Thus, results suggest that lightning production of NOx inside a thunderstorm can lead to significant impacts on surface ozone concentrations in the tropics. Enhancement in NOx at the surface after thunderstorm activity is much greater in premonsoon periods compared to the monsoon period.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Nitric oxide (NO) is the most important electric-discharge-produced molecule, primarily because it facilitates chemical reactions in the troposphere and stratosphere that determine the concentrations of ozone (O3) and hydroxyl radical (OH). Bond et al.  compared the geographical and seasonal production of NOx by lightning to other sources like anthropogenic activity, biomass burning, and soil emissions. Their results indicate that production of NOx by tropical lightning is significant throughout the year. In the troposphere, lightning and aircraft exhaust contribute to ambient NO levels, which can either increase or decrease O3 levels depending on the local conditions and level of NOx (NOx = NO + NO2). Lightning channels in air generate NO primarily as a trace gas with considerably less NO2. NO once present, is always accompanied by NO2 produced from oxidation of NO. In the sunlit troposphere, NO2 is photolyzed (broken into NO and O by light photons); subsequent chemical reactions and the photolyzation of NO2 establish a balance between NO and NO2 on a time scale of minutes. In the 1970s, it was established that NO produced by atmospheric electric discharges can indeed play an important role in the chemistry of the Earth's atmosphere, particularly, as noted earlier, in regulating the concentrations of O3 and OH [Crutzen, 1970, 1973; Chameides and Walker, 1973]. The first direct observations of enhanced NO concentrations in the vicinity of lightning flashes were made by NOxon [1976, 1978] who used a solar absorption spectrometer to measure NO2 concentrations below the cloud bases of lightning producing and nonlightning producing storms. He observed increases in NO2, about 2 orders of magnitude above ambient levels and as high as 100 ppbv (parts per billion by volume). Evidence of NOx production by lightning was given by airborne measurements in and near mature thunderstorms. Lightning can lead to significant increases in NOx in the middle and upper troposphere and contribute significantly to columnar abundance of NO2. [Choi et al., 2005; Zhang et al., 2000, 2003; Tie et al., 2002; Bond et al., 2001; Pickering et al., 1998; Delmas et al., 1997; Ridley et al., 1996; Beirle et al., 2006; Martin et al., 2006]. Schumann and Huntrieser  reviewed the global lightning-induced nitrogen oxides source. In the atmospheric boundary layer ozone is generated and removed by catalytic and photochemical reactions involving CO, VOCs (Volatile Organic Compounds) and NOx. The major oxidants ozone and hydroxyl (OH) are not emitted directly into the atmosphere and their concentrations are determined by complex interplay between dynamics and chemistry which is function of NOx, CO, H2O and hydrocarbons. However, NOx and ozone due to lightning activity for the Indian region has not been studied extensively. It may be noted that monthly average lightning flash counts over Gangetic West Bengal are maximum in the month of May [Kandalgaonkar et al., 2006] which is normally the hottest month of the year in most parts of India. However, the spatial distribution pattern of ozone and its precursors over the Indian subcontinent during this season is mainly controlled by the dynamics and in some cases by episodic events. In this study the data of surface NOx and ozone on thunderstorm events have been analyzed along with electric field measurements in premonsoon and monsoon seasons.
2. Methodology and Data
2.1. Observation Site Location, Wind, NOx Emissions
 The observational site is situated at the northwest region of Pune city (18.54°N, 73.81°E), which is located at a height of 559 m above sea level on the leeside of western Ghats. The site is influenced by the Indian southwest monsoon because its location is close to the Arabian Sea. The site receives southwesterly winds pattern during the southwest monsoon season (June–September), and northeasterly winds during winter (October–January) and some periods of summer (February–May). Figure 1 is a map of India showing the observation site and climatological winds at 925 hPa for a premonsoon month (May) (Figure 1a); and monsoon month (July) averaged for 30 years from 1971 to 2000 taken from NCEP wind reanalysis (Figure 1b).
Figures 2a and 2b show the high-resolution gridded (0.25°) anthropogenic NOx emissions over Indian region in 2008 from all the sources and from fossil fuel, respectively. Total NOx emission for India is found to be around 6583 Gg/yr in 2008 as compared to 5723 Gg/yr for 2005 with an increase of around 15% (860 Gg). The spatial pattern shows that maximum NOx emission is found to be of the order of 80–200 Gg yr −1 box−1 over western, Indo-Gangetic Plain (IGP) area, eastern, southeastern regions and eastern parts of central India as well. Fossil fuel consumption is major contributor to emissions in Indian cities. Coal used in thermal power stations, transport sector, and industrial practices are the dominant contributors to the above regions, followed by biofuel used for residential purposes in the above discussed regions. The contribution of fossil fuel is found to be around 90% with an increasing trend over many parts of the Indian geographical region. Anthropogenic NOx emission (∼54 Gg/yr) in Pune (the eight largest metropolis in India, (and the second largest city in the state of Maharashtra) is mainly from the transport sector, followed by fossil fuel used in the industrial and residential sectors. An emission of the order of ∼39 Gg/yr from transport sector is the major contributor to total NOx emission in Pune city followed by industrial (∼12 Gg/yr) and residential emission (∼3.5 Gg/yr). There has been a rise of 33% in NOx emission since 2005, and future increases are anticipated. Sources of NOx emission from the transport sector is scattered over city, accompanied by a large number of industrial units [Ali et al., 2009; Beig and Brasseur, 2006; Sahu, 2010].
2.2. Measuring Instruments
 The electric field is measured with an AC field mill kept in a pit with its sensor flush with the ground. It can measure electric field in the range of ±12.5 KV m−1 and has a response time of 0.1 ms. The field mill is calibrated in the laboratory by placing a flat circular plate 10 cm above and parallel to another similar grounded plate kept level with the field mill sensors and raising it to different potentials. Its response is found to be linear with the electric field of up to ±12.5 KV m−1 applied between the sensors of the field mill and the plate placed above them. Signals from the field mill are amplified and fed through coaxial cables to a data logger which digitizes the signals using a 12-bit analog-to-digital converter for recording and storage of the data at a frequency of 10 Hz. Details of the equipment are given by Pawar and Kamra [2002, 2004]. Total flash rate is counted from the electric field record; a field change of at least 600 Vm−1 occurring in a period of 2 s being taken as a lightning-induced change [Pawar and Kamra, 2002, 2004].
 The analyzers used in the present study for the measurements of different pollutants are U.S. EPA (United States Environmental Protection Agency) approved. Ozone is measured using an analyzer (O342M, Environment S.A., France) based on the UV absorption technique, using a wavelength of UV radiation at 253.7 nm. The analyzer automatically incorporates the corrections due to changes in temperature and pressure in the absorption cell and drift in the intensity of UV Lamp. The lower detection limit of the analyzer is 1 ppbv with noise of about 0.5 ppbv, under 10 s response times. The instrument recorded the data in the range of resolution 1 min to 24 h. In these study hourly averages of concentration is used. The NOx is measured with the Horiba, Japan made APNA 365 analyzer based on the technique of chemiluminescence. Most of the NOx measurements depend upon some kind of surface conversion device to convert NO2 to NO (Molybdenum in the present case), which can then be detected by chemiluminescence. It has been well established now that in the molybdenum converter instruments other gas phase nitrogen containing compounds are converted by molybdenum oxide catalysts to NO and therefore can be reported as NO2 [Winer et al., 1974]. Therefore, many studies [Lamsal et al., 2008, 2010; Dunlea et al., 2007] have shown that large positive bias can exist in the measurements of NOx done by such instruments. Even though, in our instrument no attempt has done to find a bias in the measurements, however as the bias is mainly due to the concentrations of nitrogen containing compounds, which may not vary on day to day basis. Therefore, we think that bias will be same for thunderstorm and nonthunderstorm days; moreover most of the thunderstorms have occurred during same time of the day. Therefore, we feel that biases in the measurements may not affect much on the conclusions of this study. This instrument records the data from the intervals of 3 min to 3 h. In this study data is sampled at 3 min interval and averaged for 1 h. The lower detection limit of the analyzer is 100 pptv. However, most of the observed values are found to be well above this limit, which provides very large signal-to-noise ratio for corrected measurements. On thunderstorm and nonthunderstorm days we analyze surface measurements of electric field, NOx and O3, which show the influence of thunderstorm activity on the above parameters.
 The region around Pune generally experiences thunderstorms during the premonsoon season (March–June) and on initial days of postmonsoon (September–November) months, whereas thunder is rarely observed during the monsoon period (July–August). The development of isolated thunderstorms in this area in the premonsoon season is generally due to instability of the lower troposphere created by surface heating. Base height of these thunderstorms is generally 1 to 2 km. Thunderstorms in this region are generally stationary.
3.1. Thunderstorms During the Premonsoon
 Over the Pune region thunderstorms are frequently observed during April to May. Table 1 shows the detail statistical information about total number of events studied during the premonsoon and monsoon in the period of 2005–2008. We analyzed a total of 60 events of thunderstorms of which, 45 cases belong to premonsoon season and 15 to monsoon season. In this study we analyzed electrical parameters (lightning flash rate, electric field) with the chemical species (NOx, O3) during the thunderstorm activity using the thunderstorm and nonthunderstorm day data. Figures 3a and 3b show the lightning flash rate per five minutes and electric field during thunderstorm observed on 3 May 2005, respectively. Lightning flash rate has been derived from electric field data. Figures 4a and 4b show diurnal variations of NOx and ozone observed on 3 May 2005 as well as 5 days before and after the thunderstorm day (nonthunderstorm days). The thunderstorm developed southeast of the observatory at about 16:30 h Local Time (LT). As shown in Figure 3b the electric field became positive and increased rapidly around 17:30 and lightning activity began. Intense lightning activity was observed for 2 h from 17:10 to 19:10 with decreases in frequency thereafter. The lightning activity stopped completely by 19:15 h LT. It is well known that the lightning is a major source of NOx in the middle and upper troposphere and contributes significantly to abundance of NO2 [MacGorman and Rust, 1998; Choi et al., 2005; Beirle et al., 2006; Martin et al., 2006; Schumann and Huntrieser, 2007]. However, there are only a few studies about surface variation of NOx and ozone associated with thunderstorms [Betts al., 2002; Ott et al., 2010; Minschwaner et al., 2008]. The lightning frequency and electric field records shown in Figures 3a and 3b clearly suggest that the dissipation stage of this thunderstorm had started at about at 19:00 h. The concentrations of NOx and ozone observed on nonthunderstorm days are considered as representative of the background levels. As shown in Figure 4a NOx concentration (14 ppb) after 2–3 h (21.00) of peak lightning activity is more than the twofold compared to the nonthunderstorm days. This increase in NOx suggests the transport of NOx from the upper or middle troposphere by downdraft of the thunderstorm during its dissipation stage. Figure 4b shows the variation of ozone on 3 May 2005 thunderstorm day with 5 days before and after the thunderstorm day (nonthunderstorm days). A study by Minschwaner et al.  shows enhancements in ozone between about 3 and 10 km altitude within an electrically active storm in central New Mexico. However, our analysis shows a decrease in ozone following the dissipation stage of thunderstorm (Figure 4b). Several studies have shown that the impacts of NOx on ozone are complex and nonlinear [Lin et al., 1988]. The decrease of ozone following the thunderstorm on 3 May 2005 can be attributed due to the “titration reaction” which became dominant only after the NOx titration threshold is reached. Ozone production is typically associated with hydrocarbon oxidation that produces HO2 or peroxy radicals (RO2) such that
Combined with reactions (R3) and (R4), these reactions convert NO to NO2 without the consumption of ozone, resulting in net ozone production.
O3 production increases linearly with hydrocarbon concentrations but varies inversely with NOx concentrations (hydrocarbon-limited regime) because the O3 production rate is limited by the supply of hydrocarbons. The dependence of O3 production on NOx and hydrocarbons is very different between the two regimes. Ozone concentrations (ppbv) simulated by a regional photochemical model is a function of NOx and hydrocarbon emissions [Sillman et al., 1990]. In the NOx-limited regime, hydrocarbon emission controls are of no benefit for decreasing O3. In the hydrocarbon-limited regime, NOx emission controls cause an increase in O3. O3 production varies linearly with the NO concentration but is independent of hydrocarbons (NOx-limited regime) because the O3 production rate is limited by the supply of NOx. On a thunderstorm and nonthunderstorm day peak of O3 occurs during afternoon hours (12:00–15.00) because of strong solar radiation (Figure 4b). After 2–3 h of thunderstorm activity (21.00 h) NOx slowly increases and reaches concentration (14 ppb) above the titration threshold level and starts to destroys the O3 (∼15 ppb) via titration reaction. Peak of NOx and dip of O3 at 21:00 h on thunderstorm day is clearly seen in Figures 4a and 4b because of thunderstorm and lightning activity. However, the threshold limit is not fixed and varies from region to region. Based on the long-term data and their analysis as reported earlier from the same station [Beig and Brasseur, 2006; Beig et al., 2007], the threshold limit is estimated to be between 12 and 20 ppb [Beig et al., 2010].
Table 1. Present Concentrations of NOx and O3 for Premonsoon and Monsoon Thunderstorm Events
Values outside parentheses are average values; values within parentheses are standard deviation.
NOx above threshold
NOx below threshold
NOx above threshold
NOx below threshold
 We discuss one more case of premonsoon thunderstorm observed on 3 June 2008. This thunderstorm developed at about 16:00 h LT local time and lasted for about 2 h; the flash rate along with 5 min average of electric field is shown in Figures 5a and 5b. Lightning flash rate increased very rapidly and reached 120 flashes per 5 min during the active stage of the thunderstorm and then decreased gradually to less than one flash per minute at about 18:00 h. As shown in Figure 5b with the onset of the thunderstorm, the electric field increased to 2 kV/m and lightning activity started at about 16:20 h. In Figures 6a and 6b we have plotted diurnal variations of NOx and ozone observed on 3 June 2008 as well as 5 days before and after the day of the thunderstorm (nonthunderstorm days). As shown in Figure 6a, NOx shows considerable increase at 21:00 h. This considerable increase in NOx observed on 3 June 2008 at 21:00 h is clearly due to transport of NOx from the upper or middle troposphere with the downdraft of the thunderstorm during its dissipation stage. Figure 6b shows the variation of ozone on 3 June 2008 and 5 days variations of ozone before and after thunderstorm day. During the dissipation stage (20:00–21:00 h) of the thunderstorm, the surface NOx (17 ppb) increases sharply and at the peak level becomes four times more than the normal fair weather days (Figure 6a). The increase at peak hours is found to be from 4 ppb for nonthunderstorm days (fair weather days) to 17 ppb during thunderstorm hours. The rate of increase is high in the magnitude of NOx concentration which is close to titration threshold limit. As a result the ozone level drops down due to the titration reaction. We did not find any enhancement in ozone during the active period of the thunderstorm (Figures 4b and 6b) as also reported earlier by Minschwaner et al. . In fact it is observed that in most of the cases, ozone showed low concentrations during the active period of thunderstorms. This decrease in concentration may be due to reduced solar radiation because of the presence of thunderclouds. However, it is observed that there is a reduction in ozone during the dissipation stage of the thunderstorm as seen in Figures 4b and 6b. During premonsoon season 10 out of 45 cases shows that there is large enhancement in NOx concentration above the titration threshold during or after the dissipation of the thunderstorm and reduction in surface ozone concentration.
3.2. Thunderstorms During Monsoon
 Indian summer monsoon onset occurs in the first week of June over the Pune region, and thunderstorms are rarely observed during July to September. About 15 cases of thunderstorms are observed during this period. The thunderstorms formed in this period are low in intensity because of weak heat convection during the monsoon period. Figure 7a shows the lightning flash rate per five minutes observed on 13 June 2005 and Figure 7b shows the electric field recorded during a monsoon thunderstorm observed on the same day. The thunderstorm developed southeast of observatory around 14:40 h. The electric field became positive and increased rapidly at about 14:55 h and lightning activity began. Intense lightning activity was observed for 2 h from 15:30 h to 17:30 h, and then stopped completely at about 18:00 h. Diurnal variations of NOx and O3 observed on 13 June 2005 along with 5 days before and after the thunderstorm day (nonthunderstorm days) are shown in Figures 8a and 8b. On 13 June 2005, thunderstorm developed little early in the afternoon, i.e., at 14:40 h. On that day during dissipation stage of thunderstorm NOx increased considerably at 17.00 h, its magnitude reached up to 7 ppb and remained less than the titration threshold limit as a result ozone increased.
 Generally, in the tropics thunderstorms develop in barotropic environments with very small vertical wind shear and therefore, the updraft and thus the growth of the storm is nearly vertical [Williams, 2001]. The accumulated condensation in the upper regions of storm frequently descends in place to create the low-level downdraft. In the extratropics, thunderstorms develop in baroclinic environments with large vertical wind shear [Williams, 2001]. Therefore most of the time dissipation stages of such thunderstorms are of shorter duration compared to the extratropical thunderstorms. Pawar and Kamra  have shown that dissipation time for thunderstorms over Pune (India) are 24% of those for the large quasi-stationary thunderstorms occurring at Florida and 55% of the convective air mass thunderstorms at New Mexico. Therefore, it is expected that the NOx transported to the surface, below tropical thunderstorms in the dissipation stage can be considerably higher than in extratropical thunderstorms. The observations discussed here show a many fold increase in NOx during the dissipation stage of thunderstorms and in some cases very large increases have been observed. On 24 April 2005 thunderstorm NOx concentration increase (70 ppb) ten times more than the background. Noxon [1976, 1978] observed that lightning flashes increases NO2, about 2 orders of magnitude above ambient levels and as high as 100 ppbv (parts per billion by volume). Bond et al.  found that the production of NOx by tropical lightning is significant throughout the year. Lightning accounts for almost all of the NOx emitted over the oceans and 50–90% of NOx emitted over some continental areas on a seasonal basis [Bond et al., 2002]. The large amounts of fresh NOx emissions do not immediately contribute to photochemical ozone production and can lead to ozone decrease by both direct titration and nighttime chemistry that further reduces ozone. The observed NOx concentration at the surface is higher than what it would be for an annual average as the summer months are the lightning-active times of the year. On 13 June 2005 (Figure 8) before and during the thunderstorm activity NOx concentration is almost same as a nonthunderstorm day, but after 1–3 h of thunderstorm activity NOx concentration increases more than twofold of nonthunderstorm day.
Table 1 presents concentrations of NOx and O3 for premonsoon and monsoon thunderstorm events. For each thunderstorm event, we have calculated the difference in NOx between the peak NOx after the storm activity on thunderstorm day and the NOx on nonthunderstorm day for the same hour after the storm (not for anytime during the day). The averages of these differences were then calculated for each season and NOx regime, and the results are presented in Table 1. The above method was also applied to the ozone measurements. For each thunderstorm event, we have calculated the difference in the O3, when the peak in NOx was observed on thunderstorm day hours and nonthunderstorm day for the same hour as taken for the storm day. It should be noted here that in Table 1 we present the differences in values of peak NOx and minimum of O3 after the storm, not for anytime during the day.
 During premonsoon, 10 out of 45 events showed that NOx (+19 (±10.35)) increases above titration threshold limit. This enhanced NOx affect the O3 and reduces O3 (−5.625 (±3.92)) concentration via titration reaction. In these 10 cases fourfold increase is found in NOx on thunderstorm days compared to that on the nonthunderstorm days NOx. On the other hand during premonsoon 35 cases show that NOx (+2.26 (±1.6)) increases but remains below titration threshold value. This enhanced NOx helps to increase the O3 (+10.87 (±3.93)) concentration. During monsoon 15 cases show that NOx increases but remains below titration threshold value. This enhanced NOx helps to increase the O3 concentration. During monsoon not a single case found which show NOx increases above titration threshold value.
 The ozone concentration show sharp decrease exactly when, NOx reaches maximum and crosses a threshold limit (Figures 4, 6, and 9). However, in some cases the ozone concentration does not show much variation from its normal values. It has been observed that premonsoon thunderstorms are more efficient in lightning production than the monsoon thunderstorms [Williams et al., 1992; Lal and Pawar, 2009]. Enhancement in NOx after thunderstorm activity is much greater in the premonsoon period compared to that in the monsoon period (Table 1). Lightning activity in a thunderstorm is closely related to storm dynamics. In the severe thunderstorms lightning frequency is high [Williams, 2001] and therefore more number of electric discharges gives the large amount of NOx [Zhang et al., 2000, 2003].
 Present work shows the effect of lightning activity on surface NOx and O3 over a tropical station during premonsoon and monsoon seasons. Results indicate that if NOx concentrations cross the titration threshold level it reduces the ozone concentration. However, if enhanced NOx levels remain below the titration threshold level then it either contributes to elevated ozone levels or does not make any impact on surface ozone, depending on meteorological conditions.
 The author Vidya S. Pawar would like to acknowledge the motivation and support from P. N. Sen and P. Pradeep Kumar, Department of Atmospheric and Space Sciences, University of Pune, India.