4.1. Yearly Average Values and Standard Deviations for PM10, PM2.5, SO2, NO2, and CO concentrations
Table 3
Yearly average values of PM10, PM2.5, SO2, NO2, and CO concentrations by CAQMS.
Station | PM10 | PM2.5 | SO2 | NO2 | CO |
Avg (µg/m3) | Stdev | Avg (µg/m3) | Stdev | Avg (µg/m3) | Stdev | Avg (µg/m3) | Stdev | Avg (mg/m3) | Stdev |
Industrial | 67.64 | 40.77 | 37.08 | 27.91 | 16.94 | 11.69 | 20.27 | 10.39 | 1.41 | 0.84 |
traffic | 58.67 | 26.05 | 23.55 | 17.39 | 13.47 | 9.74 | 17.41 | 4.05 | 0.99 | 0.57 |
dumpsite | 51.68 | 37.54 | 25.30 | 22.09 | 15.26 | 6.60 | 16.63 | 2.51 | 0.54 | 0.36 |
harbour | 49.21 | 36.33 | 26.44 | 22.73 | 5.02 | 14.98 | 16.92 | 3.94 | 0.96 | 0.55 |
residential | 40.34 | 22.12 | 23.98 | 13.15 | 4.14 | 10.70 | 9.32 | 1.50 | 0.71 | 0.29 |
Table 3 presents an overview of annual average concentrations and standard deviations for PM2.5, PM10, SO2, NO2, and CO emissions at five monitoring stations within Chennai city. The industrial site exhibits higher concentrations of PM10 (67.64 µg/m³) and PM2.5 (37.08 µg/m³), indicative of particulate emissions from manufacturing processes and material handling. Elevated levels of SO2 (16.94 µg/m³), NO2 (20.27 µg/m³), and CO (1.41 mg/m³) suggest contributions from industrial combustion and emissions, aligning with previous studies(Bhanarkar et al. 2005; Panda and Shiva Nagendra 2018) .
At the traffic site, high concentrations of PM10 (58.67 µg/m³) and PM2.5 (23.55 µg/m³) are observed, attributed to vehicular emissions and road dust. Moderate levels of SO2 (13.47 µg/m³), NO2 (17.41 µg/m³), and CO (0.99 mg/m³) indicate the influence of traffic-related sources, including exhaust emissions (Badami 2005; Manju et al. 2018) .
The dumpsite area experiences relatively lower concentrations of PM10 (51.68 µg/m³) and PM2.5 (25.30 µg/m³) compared to industrial and traffic areas. However, significant levels of SO2 (15.26 µg/m³), NO2 (16.63 µg/m³), and CO (0.54 mg/m³) are linked to the burning of waste materials and localized emissions(Peter et al. 2018). where air dispersion predominantly impacts SO2 and CO levels, can be attributed to localized emissions of PM10 and PM2.5 from vehicle and goods handling in the market area, while SO2 and CO emissions from shipping activities are more effectively dispersed by sea breezes, leading to their dilution over a wider area.
The harbour area exhibits concentrations of PM10 (49.21 µg/m³) and PM2.5 (26.44 µg/m³) similar to the industrial site. Notably, lower levels of SO2 (5.02 µg/m³) and CO (0.96 mg/m³) reflect the dispersion effects of sea breezes and the maritime environment. The average NO2 concentration (16.92 µg/m³) indicates a balance between maritime emissions and dispersion mechanisms(Saxe and Larsen 2004; Sorte et al. 2020a). The dispersion predominantly impacts SO2 and CO levels, can be attributed to localized emissions of PM10 and PM2.5 from vehicle and goods handling in the market area, while SO2 and CO emissions from shipping activities are more effectively dispersed by sea breezes, leading to their dilution over a wider area.
In the residential area, PM10 (40.34 µg/m³) and PM2.5 (23.98) levels are moderate, originating from combination of traffic, local sources, and domestic activities. SO2 (4.14 µg/m³), NO2 (9.32 µg/m³), and CO (0.71 mg/m³) concentrations ensue from a mix of sources, including vehicular traffic, cooking and local waste burning.
4.2 Statistical distribution of Air Quality Data (PM 10 , PM2.5, SO2, NO2, and CO concentrations) at contrasting urban sites
As demonstrated in the previous section, the mean pollutant concentrations exhibit significant variations among the studied sites, as detailed earlier in the paper. The industrial site consistently registers the highest levels for PM10, PM2.5, SO2, NO2, and CO, highlighting a distinct air quality profile characterized by elevated pollution levels. In contrast, the residential area consistently displays lower mean concentrations for these pollutants, illustrating a contrasting scenario with generally lower pollution levels. The variability in the industrial site is also evident in the highest standard deviations across all pollutants, indicating greater dispersion in air quality measurements. This heightened variability underscores the complexity and fluctuation in pollution levels at the industrial site. Similarly, the standard deviation in the residential area consistently demonstrates lower values, implying more stability in pollutant levels and a more controlled and predictable air quality environment.
The histogram of pollutants and skewness and kurtosis values are shown in Fig. 2 and Table S3(supplementary data), respectively. They offer insights into the distribution of major air pollutants in Chennai.
PM10 concentrations are widespread at the industrial site and the levels are found to be rather high. This indicates that there are many sources of particulate matter such as emissions from manufacturing processes, machinery, and other industrial activities (Ghosh et al., 2018; Han et al., 2020). In contrast, at the traffic site, where vehicular presence is dominant, the histogram leans towards higher values while being comparably narrower than that of the industrial site. This tendency reflects the likelihood of significant emissions from vehicles and the consequent suspension of fine particles due to traffic movement (Ghosh et al., 2018; K A O and Friedlander 1995). At the harbour area, a narrower distribution skewed towards lower concentrations is evident, highlighting the potential dispersion effects of sea breezes (Saxe and Larsen 2004; Sorte et al., 2020). The observed distribution pattern in the dump site is similar to that of the harbour site. While earlier studies have suggested that the dump site's histogram can potentially exhibit a bimodal distribution, reflecting both moderate and high concentrations attributed to dumpsite-related activities, it is plausible that the variance can be influenced by the predominant wind direction, differing from the monitoring site's location. In the residential area, a symmetrical distribution with moderate concentrations emerges, possibly influenced by local sources and domestic activities(Peters et al., 2013).
Upon examining concentrations of PM2.5, it is seen that the pattern is similar to that found in PM10, albeit with slight variations due to the smaller particle sizes. These distributions provide valuable insights into the intricate dynamics of air quality, where each site's unique characteristics contribute to the overall picture of pollutant sources and their potential impact.
When examining the histogram representing NO2 concentrations, the industrial site presents a distinctive peak at higher levels. This observation finds its basis in the fact that industrial processes often encompass the combustion of fossil fuels and the operation of machinery, both of which release large amounts of nitrogen oxides into the atmosphere. The noticeable peak in the histogram accentuates the potential substantial contribution of these industrial activities to heightened NO2 levels. Furthermore, the distribution appears broader in comparison to other sites, reflecting the intricate interplay of multiple emission sources(Fouquet et al., 2007; Rojas & Venegas, 2013). In addition to direct industrial emissions, other factors like the nature of industrial activities, prevalent meteorological conditions, and the site's proximity to major road networks can influence the histogram's shape. At the traffic site, the histogram for NO2 concentrations portrays a peak at higher levels, akin to the industrial site but narrower in comparison. This concentration peak can be attributed to the dominant influence of vehicular emissions. With vehicles burning fossil fuels like gasoline and diesel, nitrogen oxides are produced as combustion by-products. Consequently, areas with intense vehicular movement experience elevated NO2 concentrations(Anttila et al. 2011; Gallois et al. 2005). An interesting trend is noted at the harbour site with a broader distribution skewed towards lower values. This pattern is probably due to the impact of emissions from maritime and shipping activities in the vicinity of the harbour. While dispersion effects caused by sea breezes contribute to the reduction of NO2 concentrations, the emissions from ship operations, marine vessels and maritime-related shipping activities can add to the environmental pollution in port areas. Emissions from marine traffic, including those from engines and exhaust systems, result in the release of nitrogen oxides into the air, thereby contributing to elevated NO2 levels (Fenech and Aquilina 2021; Isakson et al., 2001). In the residential area, the histogram exhibits a peak at lower NO2 concentrations. This peak can be attributed to a blend of factors that potentially mitigate nitrogen dioxide levels in residential neighbourhoods. Residential areas typically experience lower traffic volumes compared to bustling urban centres or industrial zones, resulting in reduced vehicular emissions. The narrower distribution shape featuring a peak in the lower concentration range indicates that the dominant sources of NO2 in the residential area are relatively moderate. While vehicular emissions and household activities can play a role, their collective impact results in a concentration range skewed towards the lower end. This interpretation provides a glimpse into the potential factors shaping NO2 distribution patterns in residential environments.
Considering levels of SO2, the industrial site has a peak at higher concentrations due to emissions from industrial activities (Bhanarkar et al., 2005; Jion et al., 2023). The traffic site also demonstrates elevated concentrations but with broader distributions. Conversely, the harbour site histogram has a narrower distribution with lower atmospheric concentrations due to the dispersion of pollutants. Dump site and residential areas also have a peak at lower values.
The histogram depicting CO levels depicts moderate to high concentrations with a broader distribution shape(Parrish et al., 1991). This observation can be attributed to the influence of industrial activities and combustion processes prevalent in this area. Industrial processes often involve the combustion of fossil fuels, such as coal, oil, and natural gas which will release noxious CO as a by-product. The machinery and equipment used in manufacturing activities also contribute to CO emissions(Tyagi et al., 2016). The skewed distribution shape at the traffic site suggests that the primary influence of vehicle emissions dominates the CO concentrations at this region. The harbour site's histogram shows moderate distribution pattern in line with local shipping and other vehicular activities. Residential areas have a narrow histogram with lower values which indicate moderate CO concentrations stemming from various sources, including cooking, generators, smoking and cleaning activities.
4.3. Spatial Variation of Pollutants (PM10, PM2.5, SO2, NO2, and CO)
The box and whisker plots in Fig. 3 illustrate a summary of hourly averaged pollutant concentrations at the five study sites, and Table S4 (Supplementary data) in the supplementary section shows the percentage of pollutant exceedances (i.e., EF > 1) at all monitoring stations. Each box represents the interquartile range (IQR) with the 25th and 75th percentiles, while the median is indicated by the dark black line within the box. The whiskers extend to show the maximum and minimum concentrations observed during the study period(Azid et al., 2015).
For the industrial site, annual box and whisker plots for PM10 and PM2.5 concentrations show a relatively wide IQR, suggesting variability in particulate levels throughout the year. The exceedance percentages for PM10 (25.75%) and PM2.5 (26.58%) indicate a substantial number of instances where these values exceed the standard. This variability could arise from fluctuations in industrial activities, encompassing manufacturing processes and material handling, coupled with the influence of meteorological conditions (Panda & Shiva Nagendra, 2018). The box and whisker plot for SO2 concentrations displays a moderate spread (exceedance percentage of 6.03%), attributed to industrial emissions and varying energy demand across different seasons. On the other hand, there are variations in both NO2 (exceedance percentage of 21.92%) and CO (exceedance percentage of 23.73%) concentrations, possibly stemming from fluctuations in industrial combustion processes and traffic volume.
In comparison to the industrial site, the annual box and whisker plots for PM10 and PM2.5 at the traffic site demonstrate a narrow IQR. However, the exceedance percentages for PM10 (14.52%) and PM2.5 (16.44%) indicate notable instances of exceeding standard values. Both NO2 (exceedance percentage of 17.53%) and SO2 (exceedance percentage of 4.66%) concentrations display a relatively confined IQR, while CO exhibits a consistent spread across the data (exceedance percentage of 20.45%).
Turning to the dumpsite, distinct patterns emerge for PM10 and PM2.5 concentrations. While PM10 exhibits a wide IQR, PM2.5 portrays a box plot with high variability. The exceedance percentages for PM10 (7.95%) and PM2.5 (11.23%) at the dumpsite indicate instances of values exceeding the standard. This variance in PM2.5 levels may stem from various combustion factors such as open burning of varying amounts of municipal waste or fires, along with meteorological influences (Peter et al., 2019). The box and whisker plot for NO2 displays a broad spread (exceedance percentage of 4.11%), while SO2 and CO exhibit moderate variability (exceedance percentages of 4.93% and 10.18%, respectively), likely influenced by waste combustion and local emissions.
In the harbour site, both PM10 and PM2.5 exhibit an annual box and whisker plot with a relatively narrow IQR, suggesting consistent particulate levels influenced by maritime activities and shipping emissions. The exceedance percentages for PM10 (12.33%) and PM2.5 (20.00%) indicate instances of values exceeding the standard.
At the residential area, the box and whisker plots for both PM10 and PM2.5 present an annual perspective marked by a narrow IQR. The exceedance percentages for PM10 (6.30%) and PM2.5 (5.75%) suggest instances of values exceeding the standard. This is attributed to a steadier pattern of particulate levels, influenced by a blend of local sources and household activities. (Bathmanabhan and Saragur Madanayak 2010; Hsu et al. 2017).
Seasonal analysis of pollutant concentrations at five sites in Chennai revealed a consistent pattern with maximum concentrations observed during the winter season, and the lowest concentrations during the monsoon season. The winter season exhibited peak concentrations for both PM10 and PM2.5, indicating potential influences from weather conditions and localized emissions(Ganguly et al. 2019). Conversely, NO2 concentrations were higher during the summer season, with moderate variability observed throughout the year. SO2 and CO levels at the dumpsite showed relatively stable concentrations across different seasons. In residential areas, NO2 exhibited moderate variability throughout the year, with elevated values particularly during the winter season. Similarly, SO2 and CO demonstrated moderate variability as well, with marginally higher concentrations during the winter compared to the summer and monsoon periods, a trend consistent with findings from previous studies(Bathmanabhan and Saragur Madanayak 2010; Hsu et al. 2017) .
For a comprehensive analysis, micro-meteorological data (Temperature and Relative Humidity (RH)) for the summer and winter seasons were analysed. The summary of meteorological parameters is shown in Table 4.
Table 4
Summary of Meteorological Parameters during summer and winter
Months | Temp(°C) | RH (%) |
Winter (19–31) (30–99) |
December | 19–29 | 45–99 |
January | 22–31 | 40–94 |
February | 20–31 | 30–88 |
Summer (24–40) (49–84) |
April | 24–35 | 63–83 |
May | 26–40 | 49–75 |
June | 24–38 | 59–84 |
From the results, it is observed that ambient temperatures ranged from 19–31°C in winter and 24–40°C in summer, with RH ranging from 30–99% in winter and 49–84% in summer. The correlation analysis indicated a medium correlation of PM10 with temperature and RH in the range of 0.29–0.48. The results of the correlation analysis, p values of PM10 with temperature and RH are presented in Table S2(supplementary data). The scatter plot depicting PM10 concentrations alongside relative humidity and temperature in the Fig. S3(supplementary data). It is observed that, PM10 concentrations exceeded regulatory limits (100 µg/m3 in industrial sites, > 75 µg/m3 in traffic sites, and 50 µg/m3 in harbour sites, dumpsites, and residential areas) when RH ranged from 71–99% and temperatures ranged from 18–25°C in the winter season. Conversely, higher pollutant levels were observed in the summer season when temperatures ranged from 24–35°C and RH was between 40–60%. The medium correlation observed between PM10 concentrations and temperature/RH underscores the multifaceted nature of air quality dynamics. While temperature and relative humidity exert influence on pollutant levels, their moderate correlation suggests that other factors also play significant roles. Notably, wind direction and speed, alongside local emissions, stand out as crucial contributors to air quality variations. Wind patterns, along with wind speed and local emissions, influence pollutant dispersal pathways and concentrations. In a subsequent section, the correlation between pollutant concentrations and wind speed and direction is examined.
4.4. Temporal variation of PM10, PM2.5, SO2, NO2, and CO concentrations
The diurnal variations of air pollutant concentrations are shown in Fig. 4 and Fig. S4(supplementary data). PM10, PM2.5, NO2, SO2, and CO, at the industrial site exhibit distinctive patterns and influences. PM10 concentrations at this site display two significant peaks during the morning (7 am − 10 am) and evening (6 pm − 9 pm) hours. These peaks coincide with heightened traffic activity and consequent increased resuspension of road dust and vehicular emissions, as reported by previous studies (Jang et al. 2017b; Karar and Gupta 2006). The elevated evening peak may reflect a combination of industrial processes and local sources. NO2 concentrations at the industrial site exhibit a bimodal pattern, similar to that of PM10, with a higher evening peak compared to the morning. These variations align with previous findings confirming the observed trend (Jion et al. 2023). SO2 concentrations, on the other hand, maintain a relatively consistent profile throughout the day. This stability probably results from the continuous impact of industrial emissions and energy-related sources. The CO concentrations at the industrial site are notable, with gradual increases after 6 am, remaining relatively stable until a significant surge between 9 pm and 12 am. This trend may be influenced by source emission patterns and atmospheric conditions affecting pollutant dispersion.
Industrial activities, such as manufacturing processes and energy generation, contribute to elevated CO emissions during working hours, while the evening surge results from reduced dispersion. These intricate variations underline the complex interplay between local sources, traffic dynamics, industrial operations, and meteorological conditions, which significantly influence the diurnal dynamics of air pollutants at the industrial site (Sharma et al. 2019).
Unlike the industrial site, the traffic site displays a slightly different PM10 concentration pattern. Concentrations gradually rise from 9 am, peaking around 12 pm, and then remain relatively constant for the following hours before experiencing a gradual decline from 6 pm onwards. This pattern can be attributed to the site's proximity to a bustling bus depot, where traffic activity reaches its zenith during the midday period. This unique diurnal curve further supports the correlation between PM10 variation and rush hour traffic, emphasizing the significant influence of road dust and vehicular emissions (Alshetty and Nagendra 2022). The lack of a significant variation in the diurnal pattern of NO2 and CO levels at the traffic site can be attributed to its unique location near a bus depot. Unlike typical traffic sites (G. Ayoko 2005; Shiva Nagendra et al. 2020; Yu et al. 2016)where high traffic volumes are primarily concentrated during morning and evening rush hours, this site experiences consistently elevated traffic levels throughout the day due to its proximity to the bus depot. Consequently, the observed diurnal pattern differs from the conventional traffic site pattern. Contrary to past research findings, the diurnal pattern at this specific site reveals a distinctive trend. The pollutant levels gradually increase in the morning hours, followed by a slight reduction. However, high values persist until the evening, with a minor increase during evening hours and a gradual reduction after night rush hours. The levels remain low until the next morning. This atypical pattern suggests that the continuous presence of buses and traffic throughout the day, including daytime, significantly influences the pollutant dynamics at this location.
At the dumpsite, PM10 concentrations demonstrate dual peaks during the morning and late evening hours. These fluctuations in PM10 levels may be attributed to several factors, including different waste-burning activities, localized emissions, and fluctuations in traffic levels (Etyemezian et al. 2003). The prevalence of waste-burning activities and their impact on particulate matter emissions contributes to the observed variation in the morning peak, while other sources,
possibly traffic-related, may be responsible for the evening peak. When examining NO2 concentrations at the dumpsite, the diurnal pattern shows variations between morning and evening peaks. NO2 concentrations are higher during the summer season, particularly in the morning hours, and decrease during the rest of the day. This observation suggests the influence of local factors like waste combustion and meteorological conditions, on NO2 levels. In contrast to NO2, SO2 and CO levels at the dumpsite do not exhibit distinct diurnal patterns, and their concentrations appear relatively stable throughout the day. The absence of pronounced diurnal variations for SO2 and CO concentrations can be attributed to the influence of multiple emission sources, both local and regional, contributing to the overall pollutant levels at this location.
At the harbour site, PM10 concentrations display a bimodal diurnal pattern, characterized by high morning and low evening peaks. This pattern aligns with the increased activities observed in the morning hours at the harbour, such as shipping operations and cargo handling, suggesting that variations in PM10 levels are primarily influenced by harbour-related activities and local sources (Isakson et al. 2001; Sorte et al. 2020a). Regarding NO2 concentrations, the diurnal pattern at the harbour site exhibits two prominent peaks during the morning and evening hours, similar to the trend observed for PM10. This pattern may be attributed to the presence of both shipping-related emissions and local sources, such as vehicular traffic associated with harbour activities. The morning peak coincides with increased shipping operations, while the evening peak can be due to a combination of factors, including traffic-related emissions and reduced dispersion. SO2 concentrations at the harbour site also display a bimodal diurnal pattern, with higher levels observed during the morning hours. This pattern aligns with increased maritime activities and emissions from ships. The lower evening concentrations may be associated with decreased harbour operations and changing wind patterns. In the context of CO concentrations, the diurnal variation at the harbour site indicates moderate annual variability, with slightly higher values observed during the winter compared to the summer and monsoon seasons. CO levels reflect a combination of emissions from harbour-related activities, shipping, and local traffic. The variability observed in SO2 and CO levels at the dumpsite can be attributed to the complex interplay of factors, including the influence of road traffic in close proximity to the dumpsite. The levels show a general trend of being high during daytime, with some days exhibiting a bimodal variation, while on other days, there is a gradual increase with the maximum occurring in the noon. Identifying a specific and consistent pattern proves challenging due to the combined impact of dumpsite activities and nearby road traffic. To gain a more comprehensive understanding of the influence of the dumpsite alone, future studies will involve more localized and mobile monitoring on a micro-scale. This approach will allow for a more detailed examination of pollutant variations in the immediate vicinity of the dumpsite and contribute to a more accurate characterization of its impact.
PM10, PM2.5 concentrations at the residential site also manifest a bimodal diurnal pattern, reflecting the impact of local sources and household activities. The morning and evening peaks in PM2.5 levels are presumably from a mixture of factors, including cooking activities, DG sets, and increased vehicle during rush hours. NO2 concentrations at the residential site exhibit a distinctive diurnal variation, with higher levels consistently observed throughout the daytime, from 6 am to 6 pm. This pattern suggests the influence of traffic-related emissions, including vehicular exhaust and local sources. The absence of a strong evening peak indicates continuous traffic-related NO2 emissions during daylight hours (Jion et al. 2023). Regarding SO2 concentrations, the diurnal pattern at the residential site demonstrates moderate variability, with slightly higher levels observed during the winter season compared to the summer and monsoon months. In the context of CO concentrations, the diurnal variation at the residential site reflects a mixture of local emissions and traffic-related sources. A distinct bimodal pattern is observed, with a morning peak in concentration that may be from cooking activities and an evening peak that is possibly influenced by increased traffic and cooking activities during evening hours.
The Fig. S5(supplementary data) illustrates the variations observed in PM10, PM2.5, SO2, NO2, and CO concentrations during weekends and weekdays across all five study sites. Upon conducting weekend and weekday analyses, notable trends emerged, particularly regarding PM10 and PM2.5. Across traffic, industrial, and residential areas, a consistent pattern emerged where pollutant concentrations during weekdays slightly exceeded those observed on weekends (Saturdays and Sundays). However, a subtle disparity was noted in dumpsite and harbour areas, where Sunday concentrations marginally surpassed those of Saturdays. This discrepancy could potentially be ascribed to heightened activities or operational functions specific to Sundays in these locales. Notably, in the harbour area, increased human traffic for fish procurement activities on Sundays may contribute to this phenomenon. Moreover, similar variations were observed for CO and NO2 concentrations in both dumpsite and harbour areas, aligning with the trends observed for PM10 and PM2.5.
4.5. Coefficient of divergence
Table 5
Coefficient of variation of PM2.5, PM10, NO2, SO2, and CO in all the monitoring stations
Site Combination | PM10 | PM2.5 | NO2 | SO2 | CO |
I - T | 0.29 | 0.39 | 0.32 | 0.56 | 0.37 |
I - D | 0.36 | 0.41 | 0.52 | 0.62 | 0.61 |
I - H | 0.40 | 0.41 | 0.37 | 0.62 | 0.40 |
I - R | 0.37 | 0.40 | 0.41 | 0.49 | 0.40 |
T- D | 0.28 | 0.32 | 0.56 | 0.46 | 0.51 |
T - H | 0.30 | 0.33 | 0.37 | 0.46 | 0.33 |
T - R | 0.27 | 0.31 | 0.32 | 0.38 | 0.30 |
D - H | 0.24 | 0.25 | 0.47 | 0.42 | 0.46 |
D - R | 0.37 | 0.35 | 0.52 | 0.41 | 0.41 |
H - R | 0.35 | 0.35 | 0.37 | 0.40 | 0.21 |
Note: I-Industrial site, T-Traffic site, H-Dumpsite, H- Harbour site, R- Residential site |
The Table 5 illustrates COD values for diverse combinations of two study sites. The first column denotes pairs of stations, while the subsequent five columns showcase COD values for PM2.5, PM10, NO2, SO2, and CO, reflecting spatial divergence in pollutant concentrations between the paired stations. Higher COD values, approaching 1, signify substantial divergence in pollutant concentrations, indicating variability between the two sites. Conversely, lower COD values, nearing 0, suggest greater similarity in pollutant levels, signifying spatial homogeneity (Pakbin et al. 2010).
In the current study, the combination of Industrial - Dumpsite stands out for revealing significant spatial variability in air quality. This pairing exhibits relatively elevated COD values across various pollutants, signifying notable distinctions in pollutant concentrations between industrial and dump site monitoring stations. Specifically, nitrogen dioxide (NO2) attains the highest COD of 0.52 within this combination, indicating substantial variability in NO2 concentrations. Conversely, the Harbour - Residential combination demonstrates the least variability, with carbon monoxide (CO) displaying the lowest COD of 0.21 within this pairing, implying the least variability in CO concentrations, pointing towards spatial homogeneity.
For the Industrial-traffic combination, COD values are 0.39 for PM2.5, 0.29 for PM10, 0.32 for NO2, 0.56 for SO2, and 0.37 for CO. These values suggest a moderate level of spatial variability or divergence in pollutant concentrations between industrial and traffic-related monitoring stations. Similarly, the Traffic - Dump combination exhibits considerable spatial variability, reflected by COD values of 0.32 for PM2.5, 0.28 for PM10, 0.56 for NO2, 0.46 for SO2, and 0.51 for CO. Notably, elevated COD values for NO2 and SO2 emphasize substantial differences in pollutant concentrations between traffic-related and dump site monitoring stations.
The reported values in this study surpass those documented for urban sites in Los Angeles (Pakbin et al. 2010)(mean 0.13, ranging from 0.1–0.18) and are consistent with the values reported in other studies concerning the relationship between residential and traffic sites(Lianou et al. 2007).
4.6. Pollution source identification at all the study sites
Figure 5 displays the conditional bivariate plots for the pollutants (PM2.5, PM10, NO2, SO2, and CO) at the five selected locations emphasizing the sources where the concentration is more than the 75th percentile. This method is widely used for identifying possible sources of pollution in relation to wind impacts (Carslaw and Ropkins 2012). In the industrial site, where the predominant wind direction is NW followed by E, and most industries are located in the SE, E, and NE directions relative to the monitoring station, the Conditional Bivariate Probability Function (CBPF) plots reveal distinct patterns for each pollutant at the industrial site. It is seen that the PM10, PM2.5, and NO2 concentrations are significantly high when the wind blows from the S and SE directions, coupled with wind speeds ranging from 2–6 m/s, indicative of wind-carrying emissions from industries in those areas. Likewise, maximum SO2 levels are due to wind blowing in from the S / southern direction at wind speeds of 3–4 m/s, bringing in emissions from industries situated in that direction. The unique behaviour of CO concentrations, characterized by moderately high values during low wind speeds (0–3 m/s) and maximum values when the wind originates from the S with wind speeds of 2–5 m/s, underlines the intricate interplay of local combustion and dispersal of air pollutants from emission sources by winds. This patterned response underscores the pronounced impact of local sources, particularly industries, on the observed pollutants, and further emphasizes the role of wind dynamics (direction and speed) in shaping pollution patterns at the industrial site.
The CBPF plots (Fig. 6) for different concentrations at the traffic site, located in close proximity to a bustling bus depot, offer insights into the influence of nearby activities on pollutant concentrations. There are surges in PM10, PM2.5, and NO2 levels when the wind originates from the SW direction, with speeds of 2–4 m/s. This trend can be attributed to the high traffic activity and vehicular emissions from the bus depot, which contribute significantly to the observed peaks in these pollutants. Conversely, SO2 concentrations show a distinctive pattern, peaking when the wind originates from the SE direction with wind speeds of 3–5 m/s. This divergence in the trend of air pollutant concentrations in the hotspot suggests that the sources of SO2 might differ from those of other pollutants, potentially emanating from specific localized emissions or industrial sources situated in the SE direction. At the dumpsite location, PM10 and PM2.5 concentrations peak under NW wind conditions (5–6 m/s) and S wind conditions (0–4 m/s), indicating the impact of waste burning activities on particulate matter levels. Similar trends are observed for PM2.5, with additional high levels when wind blows from the SW direction dispersing pollutants from the source of emission. NO2 levels are elevated during NE or NW winds (0–2 m/s), suggesting a mix of local and regional sources. SO2 concentrations moderately increase under NW and SW winds, possibly influenced by distant sources or dumpsite emissions. CO levels, associated with local sources like waste burning, contribute to pollution, as explained in a previous study (Owoade et al., 2021).
At the harbour location, elevated levels of PM10, PM2.5, NO2, SO2, and CO coincide with SE wind direction and wind speeds of 1.5-3 m/s. Additionally, PM2.5 and CO concentrations are elevated under different wind directions (SW, SE) as well. These attributes, local sources, and wind patterns influence pollutant levels at the harbour site. In the residential area, heightened levels of both PM10 and PM2.5 correspond to a northerly wind direction and wind speeds of 0-1.5 m/s. SO2 and CO concentrations exhibit peaks during NE and SE wind directions with wind speeds of 0.5-2 m/s. It is seen that NO2 deviates from this pattern, displaying elevated levels when the wind is from the SE and NE directions and when wind speeds range from 0.5-2 m/s. These variations underline the influence of local sources, residential activities, and specific wind directions on pollutant concentrations in the residential setting. The results obtained from this observational study of variations in air pollutant concentrations in urban hotspots clearly highlight the complex interplay between local sources of air pollution, prevailing wind speed and direction, and meteorological conditions (pressure, temperature, humidity, etc.).