Characteristics of Surface Ozone Concentrations at Stations with Different Backgrounds in the Malaysian Peninsula

One of the main challenges for countries in tropical areas is the high concentration of ozone caused by elevated levels of anthropogenic and natural ozone precursors. In this study, variations in O3 concentrations from urban, suburban and rural regions of the Malaysian Peninsula were investigated using data covering a five-year period (2005–2009) obtained from the Malaysian Department of the Environment. The principal aim of the study is to identify and describe the variations in O3 concentrations recorded at three monitoring stations with different backgrounds, namely Petaling Jaya (S2) (urban), Putrajaya (S1) (suburban) and Jerantut (S3) (rural). The study also investigated the relationship between O3 distribution and its association with nitrogen oxides (NO and NO2) and non-methane hydrocarbon (NMHC). The results showed that the highest O3 concentration was recorded in a suburban area (Putrajaya (S1) with an average daily maximum value of 60 ± 20 ppbv). The O3 concentration was influenced by the characteristics of nitrogen oxides, particularly the titration of NO. The surface O3 level was found to be influenced by solar radiation and wind direction from the busy areas, most notably Kuala Lumpur’s city centre. This study suggests that the emission of O3 precursors, particularly NOx from motor vehicles, needs to be regulated to reduce the incidence of high O3 levels in Malaysia. Keyword: Surface ozone; Nitrogen oxides; Meteorological factors; Urban; Semi-urban; Rural areas.


INTRODUCTION
Ozone (O 3 ) is considered to be secondary pollutant, photochemical oxidant and the main component of smog.It is also regarded as a crucial air pollutant in the atmosphere because O 3 is capable of causing damage to human health via respiratory disease (Ho et al., 2007;Karakatsani et al., 2010;Neidell and Kinney, 2010;Mills et al., 2011;Huang et al., 2012).In addition, exposure to O 3 can lead to a decrease in lung function (Highfill and Costa, 1995).High concentrations of surface ozone also affect vegetation and forests due to the phytotoxic nature of O 3 .Ozone concentrations greater than 40 ppbv may be harmful to the crop yield, biomass production, vitality and stress tolerance of forest trees (Fuhrer et al., 1997).Excessively high levels of O 3 may be an obstacle to a forests' capacity to seize carbon should there be an excess of carbon dioxide in the future (Karnosky et al., 2003).
Surface O 3 is formed via a complex web of photochemical reactions between precursor emissions of volatile organic compounds (VOC) and nitrogen oxides (NO x ).Complete combustion processes are the main source of NO x in the air, particularly from vehicle emissions in high traffic areas (Song et al., 2011, Chelani, 2013).VOCs are emitted directly into the atmosphere from vegetation and a variety of natural and anthropogenic sources (Sharma et al., 2000).Non-methane hydrocarbons (NMHCs) are the main group of atmospheric VOCs and a precursor to O 3 production via hydroxyl (OH) radical-initiated oxidation, and subsequent reactions with NO x (Atkinson, 1997(Atkinson, , 2000;;Tang et al., 2008).Urban air sources of NMHCs include motor vehicle combustion, power plants, industrial operations, solvent usage, landfills, liquefied petroleum gas and natural gas leakages (Placet et al., 2000;Sawyer et al., 2000;Barletta et al., 2005;Tang et al., 2007;Duan et al., 2008).
Several studies have compared concentrations of O 3 in urban, semi-urban and rural areas (Sillman 1999;Latif et al., 2012;Shin et al., 2012;Wang et al., 2012).The concentration of surface O 3 is predominantly higher in suburban areas due to the downwind plume of O 3 precursors from city centres (Monteiro et al., 2012;Wei et al., 2012).The concentration of O 3 , especially in urban and suburban areas, is contributed by the chemistry of NO x (Sadanaga et al., 2012).The titration of O 3 in areas which have a high volume of traffic, in effect, reduces the amount of O 3 in city centres.The movement of O 3 precursors to suburban areas leads to a high concentration of O 3 in these areas.Several previous studies e.g., by Trainer et al. (1987), Chameides et al. (1992), Roberts et al. (1998), Starn et al. (1998) and Wiedinmyer et al. (2001) have indicated that the amount of biogenic VOC, such as isoprene, can contribute to the quantity of surface O 3 .Nevertheless a study by Von Schneidemesser et al. (2011) indicated that the contribution of biogenic VOCs were quite low compared to the contribution of ozone from anthropogenic precursors such as NO x and VOCs which originated from motor vehicles and industrial processes.Increase in surface O 3 concentration can also be contributed by the intrusion from upper part of the atmosphere through stratospheric-tropospheric exchanges (Neu et al., 1994, Lin, 2008;Sicard et al., 2009;Hu et al., 2012).
The study by Dentener et al. (2006) attested that if the current legislated emission scenarios are to be continued, the concentration of surface O 3 in parts of the Asian atmosphere will increase considerably by the year 2030.This study aims to provide comprehensive information on the variations and the characteristics of surface O 3 at different backgrounds (urban, suburban and rural areas) in Malaysian Peninsula.The study also attempts to characterise the concentration of surface O 3 at different locations and correlate this with its precursors.

Location of Sampling Stations
The data was obtained from three air monitoring stations on the Malaysian Peninsula, namely; a suburban area, Putrajaya (S1); urban area, Petaling Jaya (S2) and rural area, Jerantut (S3) (Table 1).Putrajaya (S1) (N02°55.915',E101°40.909') is a new Malaysian administrative capital.It was declared a federal territory on 1st February 2001 and hosts the administrative offices as well as the country's Federal Government.It is located 24 km south of Kuala Lumpur and 20 km from the Kuala Lumpur International Airport, and covers an overall area of 49.30 kilometers square.Petaling Jaya (S2) (N03°06.140',E101°43.330') is located within the Klang Valley region and covers an area of 97.2 km 2 .Its location near to Kuala Lumpur's city centre and industrial areas leads to a high volume of traffic.The last air quality monitoring station selected is in Jerantut (S3), Pahang (N03°55.59',E102°22.120').Located 200 km from Kuala Lumpur, this monitoring station is situated in the middle of the Malaysian Peninsula (Fig. 1).As a background station, it is surrounded by agricultural areas and traditional Malaysian villages.In 2010, the average traffic volumes recorded over a period of 16 hours for the road near to the air monitoring stations in Putrajaya (S1), Petaling Jaya (S2) and Jerantut (S3), were 56,468, 364,029 and 1,649 vehicles respectively (Ministry of Works, 2011).

Ozone Data Collection
The air quality data collected from the air quality monitoring sites by the Department of the Environment (DOE) in Malaysia is managed by a private company, Alam Sekitar Sdn Bhd (ASMA).The O 3 concentration at the ASMA monitoring stations was measured using Teledyne Ozone Analyzer Model 400A UV Absorption.The analyzer used a system based on the Beer-Lambert law for measuring low ranges of O 3 in ambient air.The concentrations of O 3 precursors, NO x (NO and NO 2 ) were determined using the chemiluminescence measurement principle, coupled with state-of-the-art microprocessor technology for monitoring high and medium levels of nitrogen oxides (Teledyne Models 200A) while the concentration of non-methane hydrocarbon (NMHC) was determined using the field-proven Flame Ion Detector (FID) (Teledyne Model 4020).Among the three monitoring stations, NMHC was only recorded at Jerantut (S3) hence, data for this parameter was not available for Putrajaya (S1) and Petaling Jaya (S2).
For quality control and quality assurance of the air monitoring data, all monitoring instruments were calibrated regularly by Alam Sekitar Sdn.Bhd.(ASMA).For gas monitoring such as O 3 , NO x and NMHC, the instruments were scheduled to have daily auto calibration using zero air (clean ambient air, free from contaminant) and standard gas concentration.Each instrument have been calibrated manually using its individual calibration gas every two weeks.The hourly data was also checked for validation before it can be transferred to Department of Environment.The lower detection limit (LDL) for O 3 was given as 0.6 ppbv, while the LDL for NO x and NMHC were 0.4 ppbv and 0.5 ppbv respectively.

Statistical Analysis
In order to compare concentrations between stations, the distribution of air quality data were conducted using the Normal P-P Plot, the Normal Q-Q Plot and the One-Sample Kaolmogorov-Smirnov test.Analysis of variance, with the addition of the Bonferroni correction was used when the data showed a normal distribution.Analyses of variance and correlation analysis were conducted using the Statistical Package for Social Sciences (SPSS) Version 17.

Trajectory Analysis
Backward trajectory analyses were performed to access regional transport patterns and to identify possible origins and pathways of O 3 (Shi et al., 2009).The backward   (Draxler and Rolph, 2003;Rolph, 2003).The meteorological drivers used to compute the trajectories were obtained from the National Centre for Environmental Prediction (NCEP) Final Analysis (FNL) archive which is maintained by ARL.A cluster analysis was then applied to categorize the trajectories into groups of similar curvature, length and transport characteristics.The backward trajectories were determined using the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) Model introduced by NOAA through the website http://www.arl.noaa.gov/ready.html.The backward trajectories for the two main stations, namely Petaling Jaya (S2) and Putrajaya (S1), were developed for 72 h for the O 3 episode event when the O 3 concentration exceeded the limit of 100 ppbv for 1 h averaging time (the maximum value recommended by the Malaysian Air Quality Interim Guidelines, RMAQG).Three dates for highest O 3 episodes were chosen for backward trajectory for the Petaling Jaya and Putrajaya stations when the concentration of O 3 recorded was at the highest level.The selected dates for Petaling Jaya station were as follows:

Observation of High O 3 Episodes
The number of hours and days where O 3 concentrations exceed the limit of 100 ppbv (the maximum value for O 3 in ambient air as suggested by Malaysian Department of Environment) at the three selected monitoring stations are presented in Table 2.The results showed that of the three stations, suburban Putrajaya (S1) had the highest hourly O 3 concentration with 148 ppbv as recorded in 2006, followed by the urban area of Petaling Jaya (S2) which had an hourly maximum of O 3 of 140 ppbv in 2005.The number of days where the concentration of O 3 was above 100 ppbv in Putrajaya (S1) (as recorded from 2005 to 2009) ranged between 5 to 24 days per year while the number of days where O 3 was recorded exceeded 100 ppbv in Petaling Jaya (S2) ranged between 2 to 12 days per year.As expected, Putrajaya (S1) also had the highest total number of hours of exceedance annually.With regards to the concentration of O 3 for Jerantut (S3), this consistently remained below 100 ppbv throughout the entire study period.
The fact that the concentration of O 3 recorded was higher in the suburban area of Putrajaya (S1) than that for urban Petaling Jaya (S2) is confirmed by the overall daily average and average daily maximum concentrations recorded at these two stations (Table 3).For example the daily (24 h) average for the O 3 concentrations recorded at Putrajaya (S1) was 22 ± 7 ppbv while at Petaling Jaya (S2) it was 14 ± 6 ppbv.The results showed that the concentrations of O 3 are much higher in suburban areas i.e., Putrajaya (S1) station when compared to stations located in busy areas, such as Petaling Jaya (S2).Plume from the city centre, consisting of NO x, which was blown downwind to suburban areas can contribute to an elevation of O and O 3 in those areas.The low concentration of O 3 in the city centre was due to the titration processes of NO on surface O 3 as mentioned by Borrego et al. (2005), Pudasainee et al. (2006) and Costabile et al. (2007).High emissions of NO from traffic are considered to be the major reason for a low level of O 3 at Petaling Jaya (S2), a curbside station.On the other hand, the rural area Jerantut (S3) did not show high concentrations of O 3 due to the low concentrations of O 3 precursors in the background area.
The phenomenon of higher O 3 concentration recorded in the suburban areas compared to the city centre has also been noted in several other studies.For example, one such study by Xu et al. (2011) in Beijing shows that the concentration of O 3 was highest in the downwind suburban areas.Agrawal et al. (2003) also determined that areas located downwind of urban centres are associated with high concentrations of O 3 , a secondary air pollutant.NO x concentrations in rural areas are typically very low hence O 3 production is also less.However, in urban and suburban areas, the relative concentration of NO x is quite significant as a result of industrial and vehicular pollution, which easily contributes to the production of O 3 even when NMHC concentration is low (Donahue et al., 1990).

Diurnal Variations of O 3 and Its Precursors
The diurnal variations of O 3 and its precursors, i.e., NO and NO 2 , are shown in Fig. 2. The O 3 concentrations presented typical diurnal variation patterns, characterized by high concentrations mid-afternoon (13:00-15:00) and low concentrations late at night or early in the morning.Such a phenomenon is quite normal, even in rural areas like Jerantut (S3).This is due to the photo-chemical reaction of O 3 precursors, such as volatile organic carbon (VOC), with ambient air from natural sources and the long distance transport of NO x (Han et al., 2011;Pires, 2012).Additionally, surface O 3, even in rural areas, can also be attributed to by the thermal convection of upper boundary layer air masses, rich in O 3 , which are drawn down to the surface layer (Lin 2008, Notario et al., 2012).Results from this study indicated that the traffic volume influenced the concentration of O 3 precursors, i.e., NO and NO 2, at all stations, particularly in heavy traffic areas such as Petaling Jaya (S2).The concentration of NO and NO 2 was found to be at its highest during the peak hours in the morning (8.00 to 9.00 hrs) and late afternoon (17:00 to 19:00 hrs).The diurnal pattern of O 3 once again indicated that the highest concentration of O 3 was recorded in the suburban area, Putrajaya (S1), followed by Petaling Jaya (S2) and Jerantut (S3).The diurnal pattern of O 3 precursors showed that the concentration of NO and NO 2 were dominant in Petaling Jaya (S2) due to the influence of traffic.The levels of NO and NO 2 were not followed by directly high concentration of O 3 .The function of NO as a titrant (which reduces the amount of O 3 ), the movement of NO x to the areas downwind and the effect of vertical mixing due to the movement of O 3 precursors to the upper part of atmosphere, seem to determine the concentration of surface O 3 recorded, particularly in city centres such as Petaling Jaya (S2).On the other hand, the lower concentration of NO x (particularly NO) in the suburban area, Putrajaya (S1), appears to influence the high level of O 3 logged at this station.
The diurnal variations of NMHC recorded at Jerantut (S3), the rural station are shown in Fig 2 .The concentration of NMHC at this station was found to be higher in the morning (8.00 hrs) with a value of 95 ppbv and late afternoon during peak hours (19.00 to 20.00 hrs) with values of 104 and 105 ppbv, respectively.The results showed that even at the rural station, the concentration of NMHC was still effected by traffic emissions.According to Kansal (2009), ground level O 3 is created by the reaction of an excess concentration of NMHCs and NO x in the presence of sunlight.Nevertheless, the influence of diurnal variations of NMHC on the concentration of O 3 at the background station was found to be limited.This is based on the low O 3 concentration observed in the Jerantut station compared to O 3 recorded at Putrajaya (S1) and Petaling Jaya (S2) which were influenced more by NO x .

Monthly Variation of Surface O 3
The average monthly variation concentration of O 3 from January to December (2005-2009) at all selected stations are shown in the box plot of Fig. 3.At Putrajaya (Fig. 3, S1), O 3 reached its maximum monthly value in March and its minimum value in July.The monthly average concentration of O 3 was found to be strongly dependent on wind pattern.In the case of Putrajaya (S1) and Petaling Jaya (S2), the high O 3 concentrations were recorded during the northeast monsoon between December and March.On the contrary, the concentrations of O 3 at Jerantut were recorded at the highest level between May to August which was during the southwest monsoon.Based on the location of these three stations, it is possible that downwind air pollutants from Kuala Lumpur's city centre influenced the concentration of O 3 at these three sampling sites.The results in Fig. 4 show the mean of cluster trajectories for the three station's areas during the northeast monsoon (Fig. 4, a) as compared to the southwest monsoon (Fig. 4, b).The results indicate that the northeast monsoon has the capacity to bring O 3 precursors, particularly NO 2 , from Kuala Lumpur city centre to the station such as Putrajaya (S1).

Air Trajectories during O 3 Episode in 2009
To determine the characteristics of surface O 3 and its precursors during a high O 3 episode, the daily maximum and daily average concentrations in year 2009 were compared (Fig. 5).The results (see Fig. 5. S1) showed that the daily maximum concentrations exceeded 100 ppbv for seven days out of the year at Putrajaya (S1); and for ten days at Petaling Jaya (S2).On the contrary, the results (Fig. 5. S3) showed that the daily average and daily maximum at Jerantut (S3) were not of notable value because the concentration of O 3 was less than 100 ppbv.
The three highest concentrations observed at two stations, Putrajaya (S1) and Petaling Jaya (S2), during O 3 high episode in 2009 were utilized for backward trajectory modelling as shown in Fig. 6.The backward trajectories for these two main stations, Putrajaya (S1) (on 5 th February 2009, 1 st April 2009 and 13 th April 2009) and Petaling Jaya (S2) (on 7 th June 2009, 12 th June 2009 and 30 th December, 2009), were developed for 72 h.The HYSPLIT model showed that during the days that O 3 were recorded at the highest level at Putrajaya (S1), the movement of winds originated from north and north east and pass through Kuala Lumpur.The highest concentration of O 3 recorded in Petaling Jaya can be generated by both southwest and northeast monsoons.The results showed that in the city centre such as Petaling Jaya the movement of local wind will determine the source of O 3 precursors while the station in the sub-urban area such as Putrajaya the movement of wind from the city centre is very important to predict the O 3 high episode phenomenon.To support the wind trajectories from the HYSPLIT Model, wind roses for each day during the three highest episodes in 2009 were also developed (Fig. 7).The results showed that the daily wind patterns during a high ozone episode generally followed the dominant wind direction as a result of the monsoon wind.
Local causes, such sea breeze, the valley effect along with other meteorological factors like rain and temperature, may  also have influenced the movement of air in the study area.Further investigation during the highest O 3 episode showed that O 3 could have been affected by the high temperature, low wind speed and low humidity (Fig. 8).This result followed findings by Reddy et al. (2012) which indicated that O 3 concentration has a significantly positive correlation with temperature and negative correlation with wind speed and humidity.

Yearly Variation of Surface O 3
The annual O 3 concentration data for the entire Malaysian Peninsula is presented in Fig. 9. Overall results of the graphics 'box plot' showed the highest average concentrations of O 3 over the years.As can be seen from the results in Fig. 9, Putrajaya station (S1) recorded the highest variation of the three stations in the study area in 2006.The amount of UV b and nitrogen dioxide were also high at Putrajaya (S1) in 2006.Of the two other stations, the second highest concentrations of O 3 at Petaling Jaya (S2) were recorded in 2005 and 2009, while the lowest O 3 concentration was found at Jerantut (S3) station in 2007 (Fig. 9, S3).The results showed that this station had the highest concentration of O 3 in 2005 and the lowest in 2007, hence, revealing decreasing pattern.In addition, the highest O 3 concentration observed was at Putrajaya (S1) in 2005 with a value of 89 ppbv while the second highest was at Petaling Jaya (S2) in 2006 and 2008, with a value of 62 ppbv.The lowest concentration of O 3 (36 ppbv) was at Jerantut (S3) in 2005.The result showed that the concentration of O 3 recorded at its lowest concentration in 2007 at the three monitoring stations may be due to a high intensity of rain from the end of December 2006 until February 2007.These extreme precipitation events were predominantly associated with strong northeasterly winds over the South China Sea (Tangang et al., 2008).Nevertheless, the total volume of yearly rainfall recorded at the Subang Meteorological Station, a station located near to Petaling Jaya air monitoring station (S2) showed that the amount of rainfall in 2007 was not the highest recorded between 2005 and 2009 (Fig. 10).

Correlation between O 3 , and Its Precursors (NO, NO 2 , NO x and NMHC)
Overall, the concentrations of O 3 at the three stations had a negative correlation with those of NO, NO 2 and NMHC (Table 4, Fig. 11).The negative correlation between O 3 and NO x is due to the strength of titration processes by NO.This can be clearly demonstrated at stations such as Petaling Jaya.The amount of NO is followed by a high concentration of NO 2 as a result of oxidation processes.The O 3 concentrations at Jerantut station (S3) also showed a strong negative correlation (p < 0.01, r = -0.619)with NMHC (Table 4).This result indicates there is no clear contribution of natural NMHC to the formation of O 3 in the rural areas.The concentration of NMHC followed the concentration of NO due to the emission from motor vehicles and reduced the amount of O 3 .
Ozone formation is a complicated process whereby NO x reacts in the presence of sunlight.O 3 was found to be higher   at two stages of NO 2 concentrations as shown in Fig. 10.Low level of NO 2 indicates the lack of NO sources for O 3 titration while the high amount of NO 2 will produce high amount of O 3 .The main source of NO x is motor vehicle exhaust pipes in urban centres.Consequently, as expected, O 3 is formed and tends to accumulate downwind (Steer and Walton, 2003).Radiant energy (hv) from the sun initiates the chemical reactions involved in the formation of O 3 in the upper atmosphere.Radiant energy (hv) is used to break NO 2 , in the presence of UV light (λ < 424 nm), into NO and low-energy oxygen atoms; O molecules (Ghazali et al., 2010).The O then reacts with oxygen in the atmosphere to produce O 3 .In the station like Petaling Jaya, the reaction of NO with O 3 seem very dominant as indicated in Fig. 11 compared to stations such as Putrajaya and Jerantut.
The quantity of O 3 in the station such as Petaling Jaya (S2) was found to be reduced by the concentration of NO accumulated from motor vehicle exhausts as indicated by several previous literatures (Finlayson-Pitts and Pitts, 1997;Atkinson, 2000;Al-Azmi et al., 2008).According to Ahammed et al. (2006), motor vehicles, especially those which move at a low speed in the city, emit a high level of NO into the atmosphere.As a result of the subsequent reaction with O 3 , NO then oxidises NO 2 in the atmosphere.
From the results of this study, it is noticed that a negative correlation indicates a high concentration of NO 2 which in effect causes a greater decrease in O 3 concentrations.When there is a low NO 2 concentration and O 3 has a high value, the reverse process occurs.The NO 2 , in effect, liberates oxygen radicals when the lower NO 2 reaches a maximum concentration.
In this study we also tried to determine the influence of VOCs represented by the amount of NMHC recorded at background station, Jerantut.As Jenkin and Clemithaw (2000) illustrated, the formation of O 3 by NO x was dependent on the sources of the conversion of NO to NO 2 by the reaction with oxygen and the emission of species that can convert NO to NO 2 .Oxidation of NO to NO 2 also occurred through the sunlight-initiated free radical catalysed degradation of VOCs.Increased temperatures are most likely to raise the emission of biogenic VOCs (a significant O 3 precursor) where most O 3 production is confined by the availability of

CONCLUSION
The results of the study showed that suburban areas had higher levels of O 3 concentration than urban and rural areas.The results indicated that the concentration of O 3 is influenced by the amount of nitrogen oxides as a result of oxidation and photochemical processes.The quantity of NO titrates that of O 3 especially in urban areas.The movement of air from a city centre will bring high amounts of O 3 precursors, particularly NO x, to the downwind areas.Furthermore, the concentration of O 3 is likewise influenced by the intensity of solar radiation and the wind pattern.
This can be shown by the difference between the daily and monthly O 3 concentrations due to the contrasting intensity of solar radiation and the movement of plume, notably from the city centre.
The results of this study are of vital importance with regards to strategies focusing on the reduction of the level of O 3 in suburban areas.An abatement of the volume of traffic in city centres will reduce the amount of O 3 precursors in the downwind areas.The findings of this study also suggest the need to investigate other possible sources of surface O 3, such as the influence of stratospheric O 3 on surface O 3 .Determination of specific molecules of VOC must also be measured by a Malaysian authority dealing with the environment so as to determine the contribution of VOC on the formation of O 3 , particularly in suburban and rural areas.The information from this study will in effect lead to further research to investigate the impact of O 3 on human health and vegetation in selected areas of Malaysia.

Fig. 1 .
Fig. 1.A geographical map of the sampling stations located in Malaysian Peninsula.

Fig. 4 .
Fig. 4. Percentage of backward trajectories to three monitoring stations during north east monsoon (December to March) and south west monsoon (June to September). a north east monsoon, b south west monsoon.

Table 1 .
Location and description of selected air monitoring stations in Malaysia Peninsula.

Table 2 .
The number of total hours, days and the maximum concentration of O 3 > 100 ppbv recorded at the three selected monitoring stations from2005-2009.

Table 3 .
Average of daily concentration and daily maximum concentration of O 3 in the study areas.The number in brackets represents the range of the daily average and daily maximum.

Table 4 .
Correlation (r value) between parameters at monitoring stations.