Measurements of ozone and its precursor nitrogen dioxide and crop yield losses due to cumulative ozone exposures over 40 ppb (AOT40) in rural coastal southern India

Measurements of ground level ozone (O3), nitrogen dioxide (NO2) and meteorological parameters (air temperature, relative humidity and wind speed and direction) has been made for 3 years from March 2007 to February 2010 at Nagercoil (8.2°N, 77.5°E, 23 m above sea level), an equatorial rural coastal site of southern India. The monthly average of daytime maximum of O3 concentrations ranged from 28 to 50 parts per billion (ppb) with an annual average of 19.8 ppb. Similarly, monthly average of NO2 concentration ranged from 3.4 ppb to 7.7 ppb with an annual average of 5.3 ppb. The monthly variation of meteorological parameters shows the little changes being a coastal site. The estimated summer crops yield losses by 1.1–15.6 % from present O3 concentration level associated with AOT40 index 3.1–5 ppm h.


Introduction
Ground level ozone (O 3 ) is increasing at the rate of 0.5-2 % year −1 over the Northern Hemisphere due to increase in anthropogenic activities on the earth surface (Vingarzan 2004). The O 3 on which attention has becoming increasingly focused in the context of its negative impacts on human and vegetation (Ashmore et al. 2006;Amann et al. 2008). Amann et al. (2008) reported that exposure to high O 3 levels greater than 90 parts per billion (ppb) (permissible limit) hourly averages is linked to respiratory problems. Similarly, due to phytotoxic nature of O 3 greater than 40 ppb (threshold limit) has the potential to cause adverse impacts on growth and yield of agriculture and horticulture crops (Fuhrer 2009 and references therein). Further, O 3 has undesirable effects on yield quality of crops that directly affect seed and fruit chemistry as well as forage nutritive value (Royal Society 2008;Booker et al. 2009).
O 3 is the most important oxidant in the troposphere. It is formed through the photochemical oxidation of precursor pollutants such as nitrogen oxides (NO x =NO+NO 2 ), volatile organic compounds (VOC), methane (CH 4 ) and carbon monoxide (CO) in the presence of sunlight. The formation O 3 is controlled by NO x in India indicates that the O 3 concentration increases with increase of NO x concentration (Lelieveld et al. 2001). The O 3 precursor gases emission are mainly increasing due to increase of automobile sector. Streets et al. (2003) reported that the emission of NO x is increasing at the rate of 6.5 % year −1 over the Indian region. The automobile contributes highest 34 % of the total NO x emission (4.6 Tg) in India between year 2000. Similarly, Ohara et al. (2007) reported that the NO x Asian emissions increased by 3 times in the year 1980-2003 with highest increase in China followed by in India. The information on O 3 measurements is limited in India; the available studies have suggested that O 3 concentrations can reach potentially damaging levels of crops (Mittal et al. 2007;Engardt 2008;Debaje et al. 2010). Model studies by Mittal et al. (2007) and Engardt (2008) have shown that the daytime monthly mean of O 3 concentration ranges between 30 ppb and 45 ppb for the year 2000 over the Indian region and exceeds AOT40 index critical limit of 3 ppm h of 3-months growing period of agriculture crops responsible for 5 % crops yield losses by O 3 damage. Further, Debaje et al. (2010) estimated that the winter wheat and summer crops yield losses by 10 % and 15 %, respectively using AOT40 index from present measured O 3 pollution level in the rural India. Emberson et al. (2009) reported that crops respond differently to O 3 exposure in Asia (India in particular) than in USA and Europe underestimating crop yield losses occur at the present O 3 concentration. A recent synthesis of experimental data from Asia shows that today's concentrations (40-80 ppb) of O 3 cause yield losses of three staple Asian crops (wheat, rice and legumes) that vary between 5 % and 48 %, 3-47 % and 10-65 % respectively. Similarly, Feng and Kobayashi (2009) reported wheat yield losses by 9.7 % at O 3 concentration 31-50 ppb using meta-analysis study and 5 % yield losses in number of crops. Wang and Mauzerall (2004) calculated that in China, Japan and South Korea 1-9 % of wheat, rice and corn yields and 23-27 % of soybean yield were lost due to 1990 levels of O 3 and that losses may exceed 30 % by 2020 using MOZART-2 model simulation. Van Dingenen et al. (2009) estimate economic loss for four crops (wheat, rice, maize and soybean) for the world and largest losses are found in India between US$3 and US$6 billion for the year 2000 and ranking highest economic losses about 22 % (US$ 4.5 billion) followed by 21 % (US$ 4.3 billion) in China.
A few scattered O 3 measurements are available over the Indian southern coastal region (Nair et al. 2002;Debaje et al. 2003;David and Nair 2011;Nishanth et al. 2012). The daytime maximum of O 3 concentration reaches greater than 48 ppb during the summer season at Thumba (Trivandrum) and Kannur sites (Nair et al. 2002;David and Nair 2011;Nishanth et al. 2012). Similarly, the average maximum O 3 concentrations greater than 40 ppb at Tranquebar was reported by Debaje et al. (2003).
In this study, we present ground level ozone (O 3 ), nitrogen dioxide (NO 2 ) and meteorological parameters (air temperature, relative humidity and wind speed and direction) measured for 3 years from March 2007 to February 2010 (study period) for the first time in this region at Nagercoil, an equatorial rural coastal site on the Arabian Sea and Bay of Bengal, a southern extreme continental part of India. Diurnal, monthly and seasonal variations of O 3 were studied in the light of NO 2, temperature, relative humidity, wind speed and direction and prevailing atmospheric conditions. The crops yield losses estimated in summer season due to cumulative O 3 exposures over 40 ppb using AOT40 index is discussed.

General meteorology
The month April is the representative for the summer season (March-May). The bright sunshine hours are experiences during these 3 months. The sunshine duration was longer (12.74 h/day) Fig. 1 Map of India shows the measurements site-Nagercoil and some other adjacent sites used in the study during April and shorter (5.74 h/day) during November (CGWB 2008). Table 1 shows the atmospheric conditions observed during the study period at Nagercoil. The average highest maximum (minimum) temperature 31.5°C (26.4°C) in April, and lowest 30°C (24.3°C) was in January during winter season indicating less variation in temperature being a coastal site. The hot and cloudy weather with moderate rainfall prevails during pre-monsoon season (June-September). Nagercoil receives rainfall both from northeast (NE) monsoon (October-December) and southwest (SW) monsoon (June-September); however, it seems that NE monsoon is more active than the SW monsoon. The month November is a rainiest month during NE monsoon season. The total annual normal rainfall is 83 cm reported by the nearest observatory Kanyakumari (8.1°N, 77.5°E, 37 m above sea level), India Meteorological Department (IMD) (http://www.imd.gov.in). Fair weather conditions prevail in the month of February is the representative for the winter season (January-February). The surface meteorological parameters such as cloud cover and rainfall are used from Kanyakumari observatory to explore their role on O 3 formation (Debaje et al. 2003; IMD 2010).

Measurement techniques
Measurement of O 3 , NO 2 in parts per billion (ppb) and meteorological parameters were carried out using the Aeroqual Series 200 Handheld Monitor (Aeroqual Limited, New Zealand) at Nagercoil, India. The more details about the instrument are available at website www.aeroqual.com. Gas Sensitive Semiconductor (GSS) technology is used to measured the gases. The sensor head is controlled by intelligent data logger. The operating chemical parameters are O 3 (detection range 0-500 ppb) and NO 2 (detection range 0-200 ppb). The lower detection limit for O 3 and NO 2 is 1 ppbv and response time for O 3 less than 70 s and NO 2 less than 60 s. The meteorological parameters such as air temperature (°C), relative humidity (%) and wind speed (ms −1 ) and direction (16 points compass) are also measured using different sensors. The lower detection limit for temperature sensor is 0.01°C and for relative humidity sensor 1 % and response time less than 1 s. All times are given in Indian Standard Time (IST), which is ahead of the GMT by 05:30 h. The O 3 sensor was calibrated with the ultraviolet (UV) photometric O 3 analyzer (Model: O 3 42M, Environmental S. A., France) by running them together with averaging time interval of 1 h, and correlation coefficient is found to be 0.86-0.94. Similarly, NO 2 sensor was calibrated with the chemiluminescence's technique (Model: APNA 365 analyzer, Horiba, Japan) by running them together with averaging time interval of 1 h, and correlation coefficient is 0.84-0.91.
All data analyses in this study are based upon the hourly averaged of O 3 , NO 2 , temperature, relative humidity and wind speed and direction. The daily means of O 3 , NO 2 , temperature, relative humidity and wind speed and direction was calculated from the hourly average. The monthly average of O 3 , NO 2 , temperature, relative humidity and wind speed and direction was calculated from the daily mean for each month. The monthly average daytime maximum value of O 3 was calculated from averaging daily single maximum values of O 3 concentration for each month. The monthly diurnal means of O 3 , NO 2 , temperature and relative humidity was computed by averaging for all days of a month for specific hour. The annual and seasonal diurnal means of O 3 are computed by averaging for all months of a year.

Method
Crop yield losses due to cumulative O 3 exposures over 40 ppb was estimated using AOT40 index (accumulated O 3 concentration over a threshold of 40 ppb) (Mills et al. 2007). AOT40 is the hourly mean O 3 concentration accumulated over a threshold O 3 concentration of  (Engardt 2008):  (Fig 2a). The highest maximum of O 3 concentration 49.7±17 ppb (27.7±8 ppb) (average for 3 years) was observed in the afternoon and minimum of O 3 16.5±11 ppb (8.1±6 ppb) at the sunrise in April (November) (Fig. 2b). It was observed that hourly averaged daytime highest maximum of O 3 exceeds 75 ppb in the afternoon in April attributed to the high temperature and clear sky along with high NO 2 concentration (favorable conditions for O 3 formation).
The daytime maximum of O 3 concentration was observed between 14:00 h and 16:00 h, and corresponding maximum air temperature attained between 13:00 h and 15:00 h indicating peak O 3 concentration lagged behind nearly by 1-2 h from peak air temperature (Figs. 2 and 4). Debaje and Kakade (2009) reported the similar time lag from peak air temperature to peak O 3 concentration at continental rural site. Similarly, the maxima of O 3 concentration were always observed in the afternoon in all seasons similar to other coastal rural sites in southern India, a characteristic of rural site (Nair et al. 2002;Debaje et al. 2003;David and Nair 2011;Nishanth et al. 2012). Figure 3a, b shows the annual average diurnal variation of NO 2 for 3 years 2007, 2008 and 2009, and diurnal variation in April and November. It is seen from the Fig. 3a that highest NO 2 concentration 6.5, 6.5 and 6.6 ppb was observed at around midnight 24:00 h and minimum of NO 2 1.7, 1.8 and 2.8 ppb in the afternoon at 15:00 h in 2007, 2008 and 2009, respectively. The maximum of O 3 concentration formed at around 15:00 h, and corresponding minimum of NO 2 observed at the same time indicate that NO 2 utilized in the O 3 formation. The significant year to year differences in annual diurnal NO 2 variation were not observed in 3 years period, however, higher NO 2 concentration was observed in 2009 than NO 2 concentration in 2007 and 2008. The maximum NO 2 concentration 8.5 ppb (4.2 ppb) was observed in the midnight and the minimum of NO 2 3.2 ppb (0.9 ppb) in the afternoon in April (November) (Fig. 3b). It is to be noted that the small upper side kink in the diurnal pattern of NO 2 were observed at around 9:00 h due to morning rush hours of traffic emission.
The high concentration of NO 2 in April was due to biomass burning and low NO 2 in November because of rainy season. The biomass burning from March to May in India results in emission of NO x and VOC lead to more photochemical buildup of O 3 in rural areas (Galanter et al. 2000). Further, the emissions of biogenic hydrocarbons (e.g. isoprene) increases with increase of temperature help produce more O 3 in summer season (Velasco 2003). Similarly, Figs. 4a, b and 5a, b show the annual average diurnal variation of temperature and relative humidity, respectively, and diurnal variation in April and November. The annual maximum temperature around 30.1°C, 30.2°C and 30.4°C was observed in the afternoon (14:00-15:00 h) and minimum of 25.5°C, 25.4°C and 25.6°C in the morning (6:00 h) in 2007, 2008 and 2009, respectively (Fig. 4a). The maximum temperature 31.7°C (29.8°C) was observed in the afternoon around 15:00 h and minimum of 27.3°C (25.5°C) in the morning 6:00 h in April (November) (Fig. 4b). The highest temperature greater than 36°C were also observed on clear sky days in April, related to highest O 3 formation. The highest relative humidity was observed in November and lowest in April, related to lowest O 3 concentration in November because photochemical formation of O 3 decreases with increase of relative humidity and overcast sky (unfavorable conditions for O 3 formation) (Debaje et al. 2003). The westerly wind of the order of 2-2.3 ms −1 was in April favor for high O 3 concentration; while northeasterly wind 1.5-1.7 ms −1 in November decreases O 3 concentration (Table 1).   Naja and Lal (1996). Similarly, annual averaged NO 2 concentration 5.3 ppb was for 3 years period with highest annual NO 2 concentration 5.8 ppb in 2009. The higher annual average O 3 concentration in 2009 than in 2007 (4.6 ppb) and 2008 (5.6 ppb) was due to the higher NO 2 concentration in 2009. The rate of increase of NO 2 is 8.6 % year −1 for the study period, which is also higher than the rate of increase of NO x 6.5 % year −1 in India between 2000 reported by Streets et al. (2003). The O 3 concentrations linearly depend on emission of NO x in the rural sites (Wang et al. 2001). It is seen from the

AOT40 index
The monthly AOT40 index value was calculated using Equation (1) from February to June in 2007, 2008 and 2009. Further, cumulative 3-months AOT40 value was calculated for different period such as February-April, March-May and April-June. Table 3 shows the summary of calculated AOT40 index values for the different period. The highest AOT40 index value was 3.1 ppm h, 2.6 ppm h and 5 ppm h in March-May 2007, March-May 2008 and March-May 2009, respectively used for estimation of crop yield losses due to O 3 damage. The AOT40 critical limits for 5 % crop yield losses for O 3 sensitive crops (watermelon, pulses (black gram (urd) and red gram (tur), green gram (moong)), cotton, wheat, onion, tomato) and moderately sensitive crops (oilseeds (groundnut, caster), potato, rice and maize) are used from Mills et al. (2007). The different AOT40 critical limits for 5 % crops losses for different crops: watermelon, pulses, cotton, wheat, onion, tomato, oilseeds,   Nair et al. (2002) b Debaje et al. (2003) c Nishanth et al. (2012) potato, rice and maize are 1. 6, 3, 3.1, 3.3, 4.1, 6, 8.9, 8.9, 12.8 and 13.9 ppm h, respectively. Large number of varities of agriculture and horticulture crops is grown over this area which is O 3 sensitive and moderately sensitive to O 3 . In addition to the above crops, millets (bajara, ragi, jowar (sorghum, sweet sorghum), sesame, bean, jute, tea, coffee, sugarcane, mangoes, bananas, orange, grape, pepper, cardamom, chilies, cashew nut, coconut and natural rubber are also grown (http://www.tnau.ac.in/tech/acpen.pdf). Table 4 shows the estimated crop yield losses for different crops using highest AOT40 index value in 2007, 2008 and 2009. The crops yield losses ranged from 1.1 % to 9.7 %, 0.9-8.1 % and 1. 8-15.6 % in 2007, 2008 and 2009 associated with AOT40 index values 3.1 ppb h, 2.6 ppb h and 5 ppb h, respectively indicate that highest crops yield losses in 2009. The higher crop yield losses were observed for O 3 sensitive crops than for moderately sensitive crops. The crops yield losses is highest for watermelon (9.7 %), followed by pulses (5.2 %), cotton (5 %) and wheat (4.7 %) in 2007. The crops yield loss is low in 2008 compared to crops yield loss in 2007 and 2009. The highest watermelon yield losses by 15.6 %, while pulses and wheat losses by 8.3 % and 7.6 %, respectively was observed in 2009. It is seen from the above estimation that AOT40 index values exceed critical limit for 5 % yield losses of crop which are O 3 sensitive during March-May indicating that the present O 3 concentration has negative effects on crop yield due to O 3 exposure in summer season.
The stomatal conductance is higher in hot and humid environments in Asia, particularly in India poses more damage to crops due to O 3 exposure than in hot and dry environments in some   (Fiscus et al. 2005). Further, crops sensitive to O 3 are at more risk due to likely co-occurrence of peak levels of O 3 and growing period of these crops from February to May. Winter wheat is especially sensitive to O 3 in India due to the likely co-occurrence of peak levels of O 3 and growing season (Emberson et al. 2003). The estimated crop yield losses by 1.1-15.6 % using AOT40 index due to O 3 damage at Nagercoil similar to other locations in India and Asia (Wahid 2006;Rai et al. 2007;Debaje et al. 2010). Wahid (2006) estimate reductions of 43 %, 39 % and 18 % in seed weight plant −1 of Pasban 90, Punjab 96 and Inqilab 91 varieties of wheat, respectively at seasonal mean O 3 concentration of 70 ppb by experimental study in Lahore, Pakistan. Rai et al. (2007) reported winter wheat yield losses by 21 % using open top chambers study for O 3 concentration of 42 ppb at Varanasi, India. The NO 2 is increasing at the rate of 8.6 % year −1 at Nagercoil due to growth of automobiles sector, where no restrictive measures on emissions of O 3 precursors (particularly NO x ) were assumed to be implemented. The elevated O 3 concentration in future has direct negative effects on crops yield production threatening the food security in India. We use the AOT40-yield response functions obtained in Europe to estimate crop production losses in the present study. The O 3 exposure-response relationships function not established in India due to limited experiments, hence, no alternative approach is possible. Our results suggest that future crop yield reductions due to O 3 damage will be large, if no mitigation efforts for O 3 precursor are implemented.

Conclusions
Measurements of O 3 , NO 2 , temperature, relative humidity, wind speed and direction was made during 2007-2010 at Nagercoil, a rural coastal site in India clearly revealed that the O 3 concentration was higher in April than in November attributed to high NO 2 concentration, high temperature and low relative humidity. The monthly average of daytime maximum of O 3 concentrations ranged from 28 ppb to 50 ppb with an annual increase rate of O 3 is 2.4 %. Similarly, monthly average of NO 2 concentration ranged from 3.4 ppb to 7.7 ppb with an annual increase rate of NO 2 is 8.6 %. The estimated summer crops yield losses by 1.1-15.6 % associated with AOT40 index values 3.1-5 ppm h. This study suggests that ambient O 3 pollution in areas of Nagercoil has potential negative impact on crop growth and yield during the summer season. We feel that more such extensive observations of O 3 and its precursor are needed to understand the effect of O 3 on crop yield losses.