Variability of the PM10 Concentration in the Urban Atmosphere of Sabah and Its Responses to Diurnal and Weekly Changes of CO, NO2, SO2 and Ozone

This paper presents seasonal variation of PM10 over five urban sites in Sabah, Malaysia for the period of January through December 2012. The variability of PM10 along with the diurnal and weekly cycles of CO, NO2, SO2, and O3 at Kota Kinabalu site were also discussed to investigate the possible sources for increased PM10 concentration at the site. This work is crucial to understand the behaviour and possible sources of PM10 in the urban atmosphere of Sabah region. In Malaysia, many air pollution studies in the past focused in west Peninsular, but very few local studies were dedicated for Sabah region. This work aims to fill the gap by presenting the descriptive statistics on the variability of PM10 concentration in the urban atmosphere of Sabah. To further examine its diurnal and weekly cycle pattern, its responses towards the variations of CO, NO2, SO2, and ozone were also investigated. The highest mean value of PM10 for the whole study period is seen from Tawau (35.7±17.8 µg m−3), while the lowest is from Keningau (31.9± 18.6 µg m−3). The concentrations of PM10 in all cities exhibited seasonal variations with the peak values occurred during the south-west monsoons. The PM10 data consistently exhibited strong correlations with traffic related gaseous pollutants (NO2, and CO), except for SO2 and O3. The analysis of diurnal cycles of PM10 levels indicated that two peaks were associated during the morning and evening rush hours. The bimodal distribution of PM10, CO, and NO2 in the front and at the back of ozone peak is a representation of urban air pollution pattern. In the weekly cycle, higher PM10, CO, and NO2 concentrations were observed during the weekday when compared to weekend. The characteristics of NO2 concentration rationed to CO and SO2 suggests that mobile sources is the dominant factor for the air pollution in Kota Kinabalu; particularly during weekdays.


INTRODUCTION
Particulate matter (PM 10 ) consists of very small solid and liquid particles suspended in air.It is of greatest concern to public health because these particles are small enough to be inhaled into the deepest parts of lung and settled there to cause adverse health effects.Particles less than 10 microns in diameter are known as PM 10 .It is a mixture of substances that include smoke, soot, dust, salt, acids, and metals.Particulate matter also forms when gases emitted from motor vehicles and industry undergo chemical reactions in the atmosphere.It is a major component of air pollution that threatens both our health and our environment.Sources of PM 10 in both urban and rural areas are not limited to motor vehicles, but also dust from construction, landfills, and agriculture, wildfires and brush/waste burning, industrial sources, windblown dust from open lands.Epidemiological studies indicated that the fine particle fractions e.g.PM 2.5 , and PM 1.0 have considerable impacts on human health even at concentrations below the Malaysia Ambient Air Quality Standards PM 2.5 of 75 3 in daily 24-hour average (Gomiscek et al., 2004).These particles when inhaled by human can evade the respiratory system's natural defences and lodge deep in the lungs.Health problems begin as the body reacts to these foreign particles.PM 10 can increase the number and severity of asthma attacks, cause or aggravate bronchitis and other lung diseases, and reduce the body's ability to fight infections (Lelieveld et al., 2015;Nirmalkar and Deb, 2015).Certain people who include children, the elderly, exercising adults, and those suffering from asthma or bronchitis are especially vulnerable to PM 10 's adverse health effects (Thurston et al., 2016).
The issue of PM 10 pollution in the urban atmosphere has received much attentions in recent years as an increasing share of the world's population lives in urban areas.Previous studies reported that traffic-related emission accounted at least 50% of the total PM emission in the urban centres (Wrobel et al., 2000).Vehicles produce exhaust and non-exhaust emissions.Aerosol exhaust emissions can be emitted directly as particles such as soot and carbonaceous aggregates that formed during the fuel combustion in the engine or formed during the emission, dilution, mixing and cooling of the vehicle exhaust gases in ambient air (Adame et al., 2014;Pérez et al., 2010).Non-combustion traffic-related PM sources are the result of the resuspension of road dust originating from the mechanical wear and degradation of tires, brakes, and pavement abrasion (Bathmanabhan et al., 2010).Therefore, the contributions to the PM 10 concentration in ambient air are segregated into two groups: (1) primary particles emitted by vehicle exhausts, and (2) nucleation particles in ambient air (Rodríguez and Cuevas, 2007).The primary particles emitted by vehicles exhausts tend to exhibit a size distribution with a nucleation mode (<30 nm) and a carbonaceous mode (50 to >100 nm).The gaseous exhaust emissions may also contribute to new particle formation in ambient air by in situ nucleation occurring sometime after emission (hours to days).This process frequently occurs in transit from the road to the urban background especially when the exhaust gas is fully diluted within the ambient air and is photo-oxidised by reactive species.
In Malaysia, the government routinely publicized the air pollution index (API) readings throughout the mainstream media to alert the public of any looming smoke haze episode.Severe haze episode often occurs during the Southeast Asian haze resulted by smoke from forest fires in Kalimantan and Sumatra (Aouizerats et al., 2015;Harrison et al., 2009).The location and background of a station, as well as wind speed, seasonal (monsoon) and weekdays-weekend variations also play important role in influencing PM 10 anomalies (Shaadan et al., 2015).Diurnal patterns, rationed between major air pollutants and sensitivity analysis, indicate that the influence of local traffic emissions on urban air quality is critical (Talib et al., 2014).A case study at three selected stations in Malaysia e.g.Petaling Jaya, Melaka, and Kuching, also revealed that the major source of air pollution over urban air is mostly due to the combustion of fossil fuel in motor vehicles and industrial activities (Norsukhairin et al., 2013).A prediction study by PCA had further identified that CH 4 , NmHC, THC, O 3 , and PM 10 are amongst the most significant factors deteriorating air quality in Malaysia (Azid et al., 2014).A critical understanding of the behaviour of major pollutants in urban atmosphere of big city is important on health concerns.It is also an essential contingency for policy implementation especially when the elevated PM concentration has caused enormous impacts on economical values (Othman et al., 2014).
The paper aims to present the descriptive statistics on the PM 10 concentration in the urban atmosphere of Sabah.Understanding the variation of PM 10 concentrations is crucial for proper health risk assessment and risk management.PM 10 was chosen because it is one of the major air pollutants listed in the Recommended Malaysian Air Quality Guideline (RMAQG) and it is also one of the five air pollutants included in the calculation of Malaysian Air Quality Index (MAPI) together with ozone (O 3 ), carbon monoxide (CO), sulfur dioxide (SO 2 ) and nitrogen dioxide (NO 2 ).This paper also presents the results of the variability of PM 10 concentration owing to the diurnal evolution and weekly cycles of CO, NO 2 , SO 2 , and ozone, particularly in Kota Kinabalu.

MEASUREMENT AND SITE
The PM 10 monitoring sites are located at five different stations in Sabah.Fig. 1 shows the regional site map of the five monitoring stations.The location details of each station is depicted in Table 1.The town of Kota Kinabalu is the busiest urban center in Sabah, located in the eastern part of Malaysia with a population of approximately 460,000 and area 351 km 2 .Its industry is not very active and the precursor emissions are mainly due to traffic exhaust (Dominick et al., 2012;Sansuddin et al., 2011).Kota Kinabalu is at 5°58 17 N 116°05 43 E and with an altitude of 5 m.The area is topographically homogenous and the city lies on a narrow flatland between the Crocker Range to the east and the South China Sea to the west.The average temperature is 27°C with April and May are the hottest months while January is the coolest.The average annual rainfall is around 2,400 mm and varies remarkably throughout the year.February and March are typically the driest months while rainfall peaks in the inter-monsoon period in October.The wind speed ranges from 5.5 to 7.9 m/s during the Northeast Monsoon (NEM) but is significantly lower to 0. ; and (d) data acquisition system for data collection and storage.The concentration levels of pollutants in the ambient air such as sulphur dioxide (SO 2 ), nitrogen dioxide (NO 2 ), ozone (O 3 ) and carbon monoxide (CO) are measured hourly and analysed in timeweighted average (TWA) in this work.The near-surface atmospheric aerosols are regularly monitored using continuous particulate matter (PM) monitoring instrument with the adoption of Met-One Beta Attenuation Method (BAM) and manual 24-h monitoring system High Volume Air Samplers (HVAS).Instruments from the Teledyne Technologies Inc., USA, such as Teledyne API Model 100A/100E, Teledyne API Model 300/300E, Teledyne API Model 400/400E and Teledyne API Model 200A/200E are used to monitor SO 2 , CO, O 3 and NO 2 respectively.The gases SO 2 , NO 2 and O 3 are determined using the UV fluorescence method, chemiluminescence detection method and UV absorption (Beer Lambert) method respectively.

1 Variability of PM 10 in the Urban
Atmosphere of Sabah Fig. 2 shows the histogram of PM 10 concentration measured at five distinct study areas in Sabah (a) Kota Kinablu, (b) Tawau, (c) Labuan, (d) Keningau, (e) Sandakan from Jan to Dec 2012.Within the measurement period, the average hourly PM 10 measured at Tawau was 35.7±17.8 3 , Kota Kinabalu (35.0±18.4m 3 ), Labuan (34.7±18.1 3 ), Sandakan (32.8±11.5 3 ), and Keningau (31.9±18.6 3 ).Dispersity of the data was the least at Sandakan with the nominal min and max at 8.0 3 and 104 3 , respectively.The variability of PM concentrations observed at different sites are quite different where out of the five monitoring stations, Kota Kinabalu measured the highest range (KK>KEN>LBN>TWU>SDK), the highest maxima (KK>KEN>LBN>TWU>SDK), and the second highest average (TWU>KK>LBN>SDK> KEN).This could be explained by their geographical environmental conditions and population density.Undoubtedly, Kota Kinabalu has the highest PM 10 concentration because it is the state capital of Sabah with population of 462,963 for total area 352 meter-sq and population density 1,315/km 2 (Department of Statistics Malaysia, 2010).Of all monitoring stations studied in this work, Kota Kinabalu is the most urbanized area with high concentration of cities, industrial and economic activities.Population density is a fundamental amplifier of air pollution problems, and urban population growth has been an important factor in worsening air quality (Yara et al., 2017).The EPA data also indicates a strong association both between higher population densities and higher traffic densities, higher population densities and higher road vehicle nitrogen oxides (NO x ) emission intensities.The second highest population density is remarked in Labuan for 948.4/km 2 , putting it the mid of all monitoring stations.Besides, Labuan is also a federal territory of Malaysia that best known as being an offshore support hub for deepwater oil and gas activities in the region.On the other hand, Sandakan and Tawau have relatively lower population density at 179.8/km 2 and 67.06/km 2 , respectively.The lowest population density is remarked in Keningau (50.12/km 2 ) which explains its lowest hourly-averaged PM 10 concentration among all other stations.However, it has the second highest range and maxima due to its geographical location near to Kota Kinabalu.The local source from Kota Kinabalu and the advection due to wind could be one of the reasons responsible for transboundary aerosol between these two regions.
The highest maximum ~317 3 was measured at Kota Kinabalu on Jan 17 th during the morning hours 0700h-0800h, while the PM 10 concentration of other study areas remained low <80 3 on the same day and time.The exact reason for the unusual high PM 10 occurred only on Jan, 17 th from 0700h-0800h was unknown but we ruled out the regional or transboundary effects as the wind speed and wind direction was quite low at 1.7 km/h (153°N) and showed no unusual pattern throughout the past three days (see Fig. 3).Not only that, CO and NO 2 also showed no irregular pattern and fluctuated within the normal range between 0.10-1.60ppm and 0.005 to 0.020 ppm, respectively.Besides, the month of January is also not the open biomass burning period in Kalimantan and Sumatera, which is usually the main cause for low visibility and haze episode in Sabah.Therefore, the transboundary effect in this case is unlikely to cause this phenomenon.In fact, it could be due to other causes but not limited to local emissions related to traffic and open burning.Unfortunately, the local VOC data measured at the same site during the episode day was unavailable and this unusual incident was also not widely archived due to the TWA-24 h of the day was still within the EPA standard of less than   during the POST-SWM for all five study areas.Besides, no significant difference was found in the highest average PM 10 among the sites, ranging from 37 to 40 3 .After the SWM, the PM 10 levels started to decline during the north-east monsoon (NEM) and post north-east monsoon (POST-NEM) for all areas.The lowest PM 10 concentration was measured at Keningau (27±13 m 3 ) during the POST-NEM.This finding is similar to another local study in Klang Valley reported that PM 10 level was persistently high during the south-west monsoon and low during the north-east monsoon (Rahman et al., 2015;Sansuddin et al., 2011).During the northeast monsoon, the precipitation of rain will carry the pollutants to the earth; hence, reducing the level of pollutants in the atmosphere.However, during the southwest monsoon, the warmer air near the surface area rises to higher latitude, which results in turbulence and causes the pollutants to become unstable; thus resulting in a high level of pollutants in the atmosphere (Barmpadimos et al., 2011).During the wet seasons, precipitation reduces considerably PM 10 concentration by wet deposition.In addition, if it is frontal precipitation, some polluted boundary layer air is replaced by clean air from aloft.This feature may not be possible to depict by a linear relationship, but the pattern is significant and obvious for case-by-case comparisons.For example, as shown in Fig. 4d, the PM 10 concentration remained positively correlated with the dry seasons during SWM which peaked in the POST-SWM at an average value 37 3 , and negatively correlated with the wet season during NEM which based in the POST-NEM at an average value 27 3 .Nevertheless, the seasonal variation of PM 10 concentration in Sabah has no drastic change but limit to an extent of less than <10 3 for all study areas.This attribute could be due to the small seasonal variation of planetary boundary layer   (Barmpadimos et al., 2011).In contrast, the relationship between the PM 10 concentration and net irradiance is reversed for decreasing solar irradiance.All five study areas are in tropic where recei-ve fairly constant amount of solar irradiance throughout the year.Hence, it is possible that through this mechanism, relatively small difference of <10 3 between seasonal PM 10 concentration was remarked for all study areas in Sabah.5a).Fig. 5b shows that the hourly averaged CO concentration rose abruptly at 7 h and gradually decreased to the normal background concentration <0.50 ppm at 10 h and then started to hike up again at 18 h and persisted until 23 h.This bimodal distribution was also found in the hourly averaged NO 2 concentration which peaks at 7 h and 19 h (see Fig. 5d).Nevertheless, the hourly maxima for both parameters CO and NO 2 was still far beyond the hourly Recommended Malaysia Air Quality Guidelines (RMAQG) of less than 30 ppm and 0.17 ppm, respectively.In contrast, the diurnal evolution of ozone and SO 2 has no direct effect on the PM 10 trend.

2 Diurnal Evolution of PM 10 in Kota Kinabalu
The ozone profile was in accordance with the photochemical effects where it peaks at the noon hours for high intensity of UV solar light.Fig. 5c clearly exemplifies this pattern on its unimodal distribution which has a bell-curve shape peaks at 13 h.Among all parameters, SO 2 has the least variability throughout the day where the fluctuation shows no significant anomalies and far behind the hourly RMAQG of <0.13 ppm.
The two rushing hours at 7 h and 19 h are associated with the heavy-traffic hours where the number of vehicles on road is the direct reflection of the related emission of CO, NO 2 , and PM 10 .The morning peak hour is when people are going to work and sending their children to school, thus causing traffic congestion and increasing the pollutant levels.Similar findings were also reported in a study at Klang Valley where the daily concentration of PM 10 , CO, NO 2 and SO 2 was clearly recorded as being the highest during traffic congestion, particularly during the morning peak between 7:00 am-9:00 am, led to the higher amount of CO in the atmosphere (Azmi et al., 2010).In the latter part of the afternoon, PM 10 concentration peaked from 7:00 pm in line with the evening rush hour when people started returning home from work.The late evening peak can also be attributed to meteorological conditions, particularly atmospheric stability and wind speed (Rahman et al., 2015).The PM levels were lower between 0800h and 1400h because of low traffic volumes and hence low emission rates.Another possible reason could be the increase in the mixing height that favors dispersion conditions (Bathmanabhan et al., 2010).However, during the night time between 2000h and 2300h, the PM concentrations were slightly higher than the daytime (see Fig. 5).The probable reason for this may be the builtup of particles under inversion conditions.Gomiscek et al. (2004) have also reported a similar trend for three urban sites in Austria.
More specifically, we present two scenarios: (1) the morning effect and (2) the evening effect in the urban atmosphere at Kota Kinabalu on Fig. 6.The left panel Fig. 6a shows the spikes on PM levels due to the evening effect and the right panel Fig. 6b shows the spikes of the morning effect.On the evening effect, the rise of PM level started at 1800h and fall abruptly after 2000h.On the morning effect, the PM level started to rise at 0700h reaching maxima 190 3 and then fall abruptly to reaching 190 3 at 0800h.These two effects are highly related to the traffic exhaust gaseous emission such as CO and NO 2 .Nitrogen dioxides have a substantial impact on PM 10 through their atmospheric oxidation to aerosol nitrate and the CO formed from oxidation of VOCs (Wang et al., 2015).The concentrations and compositions of nitrate aerosols are determined primarily by their precursor emissions, which are mainly of anthropogenic origin.Among the major pollutants, CO, nitrogen oxides (NO x = NO + NO 2 ), PM 10 , and some types of VOCs (e.g., BTEX: benzene, toluene, ethylbenzene, and ortho-, meta-, and para-xylenes) are primarily traffic-induced, while O 3 and NO 2 are secondary trace gases formed from precursors in photochemical reactions (Yoo et al., 2015).The increase of CO was substantially driven by the road traffic emissions of the day.Similar findings are also reported in Masiol et al. (2014) where species mainly linked to road vehicle exhaust emissions such as CO, and NO 2 have a bimodal structure due to the peaks in traffic at 7-9 am and 6-8 pm.Fig. 7a and 7c shows the diurnal changes of PM 10 with comparison to CO and NO 2 , respectively.The morning peak on PM 10 level at 0700h was observed concurrently with the rise on CO and NO 2 , and the abrupt fall on PM 10 between 0900h and 1500h was also concurrent with the decrease on both parameters.Similar pattern was observed for the evening peak on PM 10 level at 1800h where the peak was corresponding to the rise on CO and NO 2 and the abrupt fall of PM 10 after 2000h was consistent to the decrease of CO and NO 2 (see Fig. 8a and 8c).This pattern was further confirmed by the correlation analysis on the scatter plot PM 10 against CO and NO 2 on Fig. 7b and 7d, respectively.Positive correlation R 2 of 0.46 and 0.23 was remarked on PM 10 against CO and NO 2 , respectively.Same goes to the correlation analysis for the evening effect where positive correlation R 2 of 0.45 and 0.42 was remarked for CO and NO 2 (see Fig. 8b and 8d).The higher correlative strength of CO further implies that PM 10 level in Kota Kinabalu is more dominantly dependent on the diurnal evolution of CO concentration.This finding was in good agreement with other studies reporting that the daily patterns of each pollutant for NO, NO 2 , CO and PM 10 peak during the morning rush hours, and fall abruptly in the noon owing to the decrease in the intensity of emission and the rapid growth of the convective boundary layers (Adame et al., 2014).
The profile of SO 2 showed no pronounced changes for both the morning and evening effect.The trend is rather irregular and has no significant pattern against the changes of PM 10 level (see Figs. 7e and 8e).Both figures show no concurrent rise of PM 10 and SO 2 , neither the significant drop of both parameters occurred at the same time.Therefore, the effects of SO 2 on the evolution of PM 10 were somewhat less reactive compared to other gaseous e.g.NO 2 and CO.A very weak correlative strength of 0.09 and 0.01 was found in the scatter plot between PM 10 and SO 2 for morning effect and evening effect respectively (see Figs. 7f and 8f), indi cating the changes of both parameters are not entirely related and significant.Another possible reason owing to the low variability of SO 2 is due to SO 2 itself has a very short lifetime and can easily transform into sulphate SO 4 through oxidation and interaction with parti cles and water vapour (Kai et al., 2007).
Ozone is a reactive oxidant gas that plays an essential role in the photochemical air pollution and atmospheric oxidation processes.Although in the upper atmosphere it acts as a barrier for UV-rays, in the troposphere it is a secondary air pollutant induced through a series of complex photochemical reactions involving solar radiation and ozone-precursors gaseous such as NO 2 (Masiol et al., 2014).When the concentration of NO 2 , CO, and PM 10 started to increase in the morning hours, surface ozone level decreased due to titration reaction with NO and the rupture of the nocturnal inversion layer (Talib et al., 2014).Figs.7g and 8g supported this findings by showing the daily pattern of ozone decreased for increasing PM 10 and vice versa.Ozone concentrations continued to increase during the noon hours due to photochemical reaction and or horizontal and vertical transport processed and reached to a daily maxima 0.035 ppm at 1330h.The decreased in ozone levels was followed by a reduction in solar radiation when reaching to evening hours at 1700h.This was also associated with the increase of NO 2 and PM 10 during the evening hours when emission of traffic started to mount up.Over the Borneo, the air mass evolution during the sunset is fast er compared to the Peninsular.In other words, solar intensity was dimmed earlier at 1800h in Borneo regions.This is the reason why the ozone level reached to the daily minima 0.005 ppm at earlier time 1900h where the solar activity induced photochemical reaction was no longer active.Hence, a negative correlative R 2 -1.35 (see Fig. 7h) and -2.77 (see Fig. 8h) was observed between the ozone and PM 10 concentration, indicating an increase on PM 10 concentration was most likely associated with the decreased on ozone levels and vice versa.In general, CO, NO, and PM 10 shows two typical patterns for urban atmosphere: two daily peaks correspond to morning mode and evening mode at 0700h and 1800h, respectively.The modes are divided by a minimum extend from 1100h to 1700h, which is assumed to the results of three factors: (i) lower emission of traffic exhaust, (ii) large availability of ozone inducing the photolysis of NO x and oxidation of CO, and (iii) enhanced dipersion potential.The last effect is more pronounced in the tropics when ozone levels are always higher during the noon hours and the local atmospheric circulation during daytime is dominated by the sea breezes, which have a key role in blowing air masses from the sea (Masiol et al., 2014).

3 Weekday and Weekend Effect of PM 10 in
Kota Kinabalu Fig. 9 shows the daily average PM 10 concentration measured in the urban atmosphere of Kota Kinabalu in Jan 2012.In this figure, two effects are highlighted: the weekend effect and the weekday effect.The weekend effect and the weekday effect are separated by the symbol "M" which denotes Monday.It is emphasized the abrupt fall in PM 10 concentration is often observed dur-  ing the weekend and gradual rise is often observed during the weekday.Within the measurement period, the PM level never exceeded 100 3 over the weekend.The PM concentration was significant low <50 3 on Sunday and slightly higher on Saturday.Meanwhile, high PM 10 concentration >150 3 was observed on two Mondays of the month on Jan, 2 nd and 16 th .A harmonic pattern was observed on the PM 10 levels where the peaks are often during the first or second day of the week and the lows are during the last two days of the week.
Fig. 10 shows the average diurnal changes of PM 10 , CO, ozone, NO 2 , and SO 2 .The weekend-weekday differences of PM 10 varied greatly at the peak hour 0700h.The highest values ~160 3 was remarked on weekday and ~40 3 on weekend (see Fig. 10a).Similar pattern is observed for CO where the highest difference is observed in the morning peak hours from 0700h to 0800h (see Fig. 10b).The increase in PM 10 concentration during morning hours as a result of direct road traffic emission is observed but this increase is more marked in the cases of CO which reflects direct exhaust emissions.Daily cycles of traffic-related exhausts NO 2 is also marked by road traffic evolution on Fig. 10d where the weekday concentration is consistently higher than that of weekend.This effect has been discussed elsewhere that the pollutant number concentration increases with road traffic intensity in the morning is due to the ultrafine particle vehicle emission and formation of new particles during dilution and cooling of the vehicle exhaust (Pérez et al., 2010).Besides, the second peak observed during the noon hours was attributed to new formation of aerosols by photochemically induced nucleation by solar radiation intensity, and by dilution of aerosols due to the growth of the mixing layer (Rodríguez and Cuevas, 2007).During the night hours, the PM 10 concentration started to decline (see Fig. 10a) owing to the decrease of the traffic congestion and also probably due to the thinning of the boundary layer depth that favors condensation and coagulation processes that sink the aerosol particles (Minoura and Akekawa, 2005).As a whole, during the weekday the levels of vehicle exhaust emission such as CO and NO 2 follow the traffic intensity with maxima during the morning hours, decreasing during the day because of dilution processes and increasing again the evening.During the night hours, the concentrations remained high owing to the reduction of the boundary layer height and probably lower wind speeds that prevent the dispersion of pollutants (Pérez et al., 2010).
The average diurnal ozone volume fraction is higher during the weekend compared to weekdays (Fig. 10c).This finding is similar to that report from Gvozdic et al. (2011).This behaviour is contrary to other air pollutants such as PM 10 , NO 2 , CO, and SO 2 where the weekend concentration of traffic pollutions were lower than weekday concentration (see Fig. 11).Ozone level increased in the afternoon during both weekdays and weekends, following the solar radiation intensity and its formation is dependent on the photochemistry process.However, during the weekends, the ozone concentration was higher.This is due to the effect of the lower general pollution levels observed during the weekends, thus reducing the interaction of ozone with other species such as nitrate or carbonaceous compounds (Pérez et al., 2010).Similar results were also reported in Qin et al. (2004) that low emission of NO during the weekend could be the reason for high ozone concentration at night.Finally, we present two types of average diurnal vari-  ations during the weekday (see Fig. 11a).Firstly, the diurnal variation with a single maxima for ozone and secondly, diurnal variations with double maxima for CO, NO 2 , and PM 10 .The single maxima for ozone occurred during the noon hours from 1300h to 1400h while the double maxima for CO, NO 2 , and PM 10 occur-red during the morning hours and afternoon hours at 0700h and 1900h, respectively.The peak of CO, NO 2 , and PM 10 concentration are in the front and at the back of the peak for ozone is a representative pattern of urban atmosphere in tropical climatic region (Gvozdic et al., 2011).It is corresponding to the increase of traffic den- sity in the morning hours from 0600h to 0800h, and afternoon hours from 1800 to 2000h.However, this pattern is somewhat less reflective during the weekend (see Fig. 11b).The bimodal distribution of CO during the weekend is still legibly observable but less significant when compared to weekday.The peak of CO during the weekend was observed at 0700h and 1900h, which is similar to that of weekday.However, the peak of PM 10 had no extreme variations during the weekend.Same pattern was observed for NO 2 levels.The hourly average PM 10 concentration was constantly capped below 50 3 (see Fig. 11b) which is an indicator of good air quality.Meanwhile, the evolution pattern for SO 2 during the weekdays and weekend have no much difference.This indicates that the traffic-related emissions is less influential on the changes of SO 2 .Therefore, the sources of SO 2 are somewhat not directly related to traffic exhausts but possibly due to the industrial emissions.This can be explained by the characteristics of NO x concentration rationed to CO and SO 2 .High CO/NO x and low SO 2 /NO x ratios values indicate mobile sources while high SO 2 /NO x and low CO/NO x ratio values typically indicate point sources, e.g. from industrial activities (Talib et al., 2014).Fig. 12 shows that the air quality level in the study area was dominantly influenced by the CO/NO 2 ratio value with values ranging from 80 to 140 for weekday and 40 to 80 for weekend.For SO 2 /NO 2 ratio, the values range from 0.10 to 0.25 for weekdays and 0.15 to 0.35 for weekend.These ratio values indicate that air pollution in the study area of Kota Kinabalu, is more seriously influenced by mobile sources rather especially during the weekday period.

CONCLUSION
In the present study, the seasonal, diurnal and weekly cycles of PM 10 concentration in the urban atmosphere of Sabah was investigated.Five distinct study areas were selected to represent the urban air of Sabah, namely Kota Kinabalu, Labuan, Sandakan, Tawau, and Keningau.The major sources of elevated PM 10 over the region was high likely related to the emissions of traffic related gaseous such as NO x and CO.On the temporal distribution, PM 10 levels remain highly concentrated during the south-west monsoon (SWM) and post south-west monsoon (POST-SWM) which is the hot and dry season in Sabah.On the contrary, the PM 10 levels were slightly lower during the north-east monsoon (NEM) which is the cold and wet season.However, the variance of change was not extremely large but relatively small difference of <10 3 between seasonal PM 10 concentration was remarked for all study areas in Sabah.
To further scrutinize the variability of PM 10 concentration due to the changes of CO, NO 2 , SO 2 , and ozone, Kota Kinabalu site was selected as it is the most urbanized central city in Sabah and has the most variability among all sites.The analysis of PM 10 concentration measured between Jan 2012 to Dec 2012 in Kota Kinabalu showed clear diurnal and weekly cycles.In the diurnal cycles, two peaks were observed at the morning and evening hours which is corresponding to the heavy traffic emissions.Similar pattern was also observable on the diurnal cycles of CO and NO 2 , which indicate that the major precursors of elevated PM 10 were the traffic and mobile sources.The profile of SO 2 , however has no much significant variability throughout the day.Ozone is photochemical reactive gaseous that peaks during the noon hours but starts to decline after the absence of intense solar radiance.Besides, the correlative analysis further highlights the positive proportionality between PM 10 and CO, PM 10 and NO 2 , while negative relationship between PM 10 and ozone.
In the weekly cycles, higher PM 10 , CO, and NO 2 concentrations were observed during the weekday when compared to weekend.The ozone concentration was somehow higher during the weekend especially during afternoon hours.It is owing to the lower general pollution levels observed during the weekends reducing the interaction of ozone with other species such as nitrate or carbonaceous compounds.Besides, two types of average diurnal variations were highlighted during the week day: (a) diurnal variation with a single maxima for ozone and (b) diurnal variations with double maxima for CO, NO 2 , and PM 10 .The peak of CO, NO 2 , and PM 10 concentration are in the front and at the back of the peak for ozone is a representative pattern of urban atmosphere in this climatic region.This pattern is less reflective during the weekend except for the unimodal distribution of ozone during noon hours.The PM 10 and NO 2 levels have no extreme variations during the weekend except for the bimodal distribution CO was still legibly observable but less significant.The evolution pattern for SO 2 during weekdays and weekend have no much difference.Finally, the characteristics of NO 2 concentration rationed to CO and SO 2 suggests that air pollution in Kota Kinabalu is more seriously influenced by mobile sources rather especially during weekday period.
the seasonal variation of PM 10 concentration measured at the five distinct study areas in Sabah from Jan 2012 to Dec 2012.The temporal distribution indicated that the PM 10 levels remain highly concentrated during the south-west monsoon (SWM) and post south-west monsoon (POST-SWM) which is the hot and dry season in Sabah.From the statistics summary, the highest PM 10 concentration measured was always

Fig. 6 .
Fig. 6.The hourly mean PM 10 concentration at Kota Kinabalu from 1-Jan to 2-Jan, 2012 (left) and 15-Jan to 16-Jan, 2012 (right).Left panel shows spikes on the evening effect and right panel shows the morning effect.

Fig. 9 .
Fig. 9. Daily distribution of hourly PM 10 concentration in Jan 2012 over the urban atmosphere of Kota Kinabalu, Sabah.Alphabet symbol M denotes Monday.

Fig. 10 .
Fig. 10.Average diurnal changes of (a) PM 10 , (b) CO, (c) Ozone, (d) NO 2 , (e) SO 2 by the day of the week.Weekday is represented by blue color line and weekend is represented by orange color line.

Table 1 .
Regional site map of the CAQM monitoring stations in Sabah.The five stations are labelled as CA0030: Kota Kinabalu; Location details of monitoring site in Sabah.
3 to 3.3 m/s during Southwest Monsoon (SWM).Tawau is on the south-east coast of Sabah surround by the Sulu Sea in the east and located at 540 kilometres south-east of Kota Kinabalu.It has tropical rainforest climate under the Köppen climate classification.The climate is relatively hot and wet with average CA0039: Tawau; CA0042: Labuan; CA0049: Keningau; CA0050: Sandakan.

Table 2 .
Summary statistics of PM 10 concentration measured in 3 at five distinct study areas in Sabah from Jan 2012 to Dec 2012.