Introduction

Air pollution is a major environmental health problem, causing millions of deaths every year around the world (WHO 2014; Cohen et al. 2017). Although air quality has significantly improved in some cities, particularly in high-income countries, many low- and middle-income countries (LMICs) still suffer from very poor air quality which increases the burden of non-communicable diseases (Robinson and Hammitt 2009; Cohen et al. 2017). Air pollution has been linked to negative health impacts including cardiovascular diseases, stroke, lung cancer, as well as chronic obstructive pulmonary disease (USEPA 2015; WHO 2016). Common sources of air pollution (heavy metals and particulate matter (PM)) in LMICs include vehicular emissions, industrial plants, power generation plants, oil burning, waste incineration, biomass burning, electric power generators, tire friction, motor oils spills, construction and demolition activities, as well as re-suspension of surrounding contaminated soils and dust (Al-Khashman 2004; Van Vliet and Kinney 2007; Amato et al. 2009). A significant increase in vehicle ownership implies that vehicle emissions (both exhaust and non-exhaust emissions) have grown significantly over time, as it is estimated that 30,000 vehicles are added yearly in Kenyan roads (UNEP 2006). There were approximately 1.5 million vehicles and 110,000 motorcycles registered in 2009 by the Kenya revenue authority. It can be estimated that there were about 3 to 3.5 million vehicles in Kenya in the years 2016 and 2017. This number is estimated to grow to 5 million in the year 2030. This increase in vehicle and motorcycle population results in more air pollution even if we consider clean vehicles produced with better technology to reduce emissions. In addition, a huge percentage of the vehicles in Kenya are used and may be more than 8 years old (Kenya Bureau of Standards 2014). This increases the burden of pollution when compared to new vehicles (Kenduiwo 2014). Together, these sources contribute to road dust heavy metal content and PM2.5 (particulate matter 2.5 μm and smaller in diameter) air pollution.

Road dust comprises of particles of various sizes; some fine particulate matter (PM2.5; < 2.5-μm diameter), respirable coarse particulates (< 10-μm diameter), and non-respirable coarse particulates (> 10-μm diameter) derived from both anthropogenic (mobile and stationary) and natural sources. Previous roadside dust studies have highlighted enrichment of road dust with heavy metals like cadmium, lead, and nickel due to traffic sources (Awadh 2013). These heavy metals have been linked to adverse health effects including respiratory system disorders, nervous system interruptions, endocrine system malfunction, immune system suppression, and the risk of cancer in later life (Atiemo et al. 2011). Road dust is important in health due to its composition. It may become airborne through suspension (or resuspension) by wind or vehicle movements (Amato et al. 2009).

While the number of studies focused on air pollution and its impacts in Africa has increased over the last decade (Victoria et al. 2014; Gaita et al. 2014; Kinney et al. 2011; Van Vliet and Kinney 2007; Boman et al. 2013; Shinggu et al. 2010), there is still deficit in knowledge in African cities compared to other locations worldwide (Kinney et al. 2011; UNEP 2016). This deficit is particularly acute given the high urbanisation and population growth rates observed in African countries (Ngo et al. 2015; Fengler 2010). These urbanisation rates, combined with unregulated traffic activities, poorly maintained vehicles, and limited air pollution control policies and or implementation, mean that air quality in sub-Saharan Africa (SSA) has deteriorated over time (Zachariadis et al. 2001; Ngo et al. 2015). Even though previous studies have addressed regional aerosols in SSA (e.g. Gaita et al. 2014), there is little reported work that has been aimed at measuring urban roadside particulate matter concentrations and road dust chemistry, particularly in East African countries. Pollution concentrations near roadways are important in SSA cities because roadways are hubs of transport, commerce, and other pedestrian activities, meaning that in these spaces, high pollutant concentrations coincide with high population exposure potential. These air pollutants tend to put a lot of pressure on the already weak medical system. Some professions, e.g. street vendors, “boda boda” riders, and traffic police, as well as residents in houses close to highways, who spend long periods of time  along or near roads and on congested roadsides are particularly predisposed to air pollutant exposures (UNEP 2006). In addition, Kinney et al. (2011) found that 49% of daily trips in Nairobi were either on foot or on bicycle placing members of the public at risk of high exposure and from respiratory diseases.

Many studies agree that the construction of highways has positive effects on population growth, by attracting immigrants, bringing an array of employment opportunities as well as housing developments (Boarnet and Chalermpong 2001; Chi et al. 2006). The expansion of Thika superhighway is no different; it has led to the radical development of homes and businesses along the highway which has resulted in population growth along the superhighway. In addition, the increasing number of vehicles and motorcycles in the country is a growing concern due to their significant air pollutants input especially COx and particulate matter (Kenduiwo 2014; CAIP 2000).

In an attempt to address the effects of urbanisation, growth, and development to the human population, this study aimed to generate new understanding of the influence of vehicle emissions to road dust chemistry and PM2.5 concentrations in urban Kenya.

Methodology

Study area

The area of study was along Thika superhighway and in Thika town. Thika superhighway is a modern highway approximately 45 km long in length. It is part of an international trunk road connecting Nairobi City with Ethiopia to the north (Fig. 1). It starts at Muthaiga roundabout, proceeds to Kasarani, Githurai Interchange, Kenyatta University, Ruiru Town, and Juja Town, and ends at the bridge near Blue Post Hotel in Thika. The land along the study site is experiencing radical developmental change as new buildings are constructed every day. Land close to Thika town is currently utilised for real-estate a shift from small-scale agriculture. Whereas close to Nairobi, the land has been utilised for construction of commercial and domestic buildings. There are a few manufacturing industries in Ruiru and in Allsops, with most of them being in Ruiru. The population of towns/constituencies along Thika superhighway according to the 2009 census are estimated to be 136,917 people for Thika, 40,446 for Juja, 238,858 for Ruiru, 191,690 for Roysambu, 194,120 for Ruaraka, and 166,044 for Starehe. The human population of all these places increases every year (Nairobi County Integrated Development Plan, 2014; Gicaci 2015; Kenya National Bureau of Statistics 2013).

Fig. 1
figure 1

Map showing the locations of sampling sites along Thika superhighway. Further details about sampling sites are given in Table 1 (Thika superhighway. (Computer generated map). Digitised from Google Earth using QGIS. (GIS software). Version 2.18.17. 2018)

Full size image

The sampling points selected for this study were Allsops (AL), Githurai (GT), Ruiru bus terminus (RR), Jomo Kenyatta University bus terminus (JJ), Commercial Street Thika (THK), and Jomo Kenyatta University as a rural background (BG) site (Table 1). Thika superhighway is one of the busiest roads in Kenya which was expanded in 2012 into a superhighway. In a study done in 2015, it was reported that the traffic volume in this highway was 1998 vehicles/h/lane during peak hours and 1300 vehicles/h/lane in off-peak hours (Kabui 2015).

Table 1 Sample locations and codes, where W designates western side of the highway while E is eastern side
Full size table

At each site (Fig. 1, Table 1), road dust samples were collected and analysed to determine the level of heavy metal contamination, relative to the background site. Furthermore, at each site, PM2.5 mass concentrations were measured to determine the contribution of road traffic to atmospheric PM2.5.

Road dust chemistry

Sampling process

The collection of road dust samples was comparable to previous work undertaken elsewhere (e.g. in Europe; Amato et al. 2009 and Africa; Abah et al. 2014). Half a meter from the pavement edge, two sub-samples from each sampling site were collected within a 1-m2 area. Dust was obtained by sweeping the pavement with a plastic brush and a plastic dustpan, put in a clean polyethylene bag which was sealed, labelled, and transported to the laboratory for analysis (Raj and Ram 2013).

Sample preparation, digestion, and analysis

To determine the level of heavy metals (Pb, Cr, Fe, Cu, Ni, and Zn), the collected road dust samples were air dried at room temperature for 5 days to a constant weight and then sieved through a 125-μm stainless steel mesh wire (Abah et al. 2014). One gram of the sample was weighed to the nearest 0.001 g and transferred to a round-bottomed flask. Ten millilitre of 50% (v/v) HNO3 was added to the sample and the sample was mixed. The flask containing the mixture was covered with a watch glass. The sample was heated to 95 °C ± 5 °C and then refluxed for 10 to 15 min without boiling. The sample was cooled and 5 ml of concentrated HNO3 was added. The cover was replaced followed by refluxing for 30 min. In case brown fumes were generated due to oxidation of the sample, this step was repeated continuously until there was no further formation of the brown fumes. The resulting solution was heated at 95 °C ± 5 °C without boiling for 2 h. The sample was cooled and then 2 ml of water and 3 ml of 30% H2O2 were added to the solution. The vessel was covered using a watch glass and returned to the heat source for warming, to start the peroxide reaction. The sample was heated until the effervescence subsided after which the vessel was cooled. One-millilitre aliquots of 30% H2O2 were added continuously with warming until the sample appearance remained unchanged. The sample was covered using an Allihn condenser to recover the vapour. The heating of the acid-peroxide digestate was left to continue until the volume reduced to approximately 5 ml. Ten-millilitre concentrated HCl was added to the sample digestate. A vapour recovery device was used to cover the sample. The sample was placed in the heating source and refluxed at 95 °C ± 5 °C for 15 min. The resulting digestate was filtered through Whatman filter paper no. 41 (USEPA 1996) and the filtrate transferred into a 100-ml volumetric flask and distilled water added to 100-ml mark. The solutions were aspirated into the AAS instrument, Shimadzu AA-6200 (experimental conditions are outlined in Table 2). The detection limits for the metals were found to be Zn 0.01459 μg/g, Cu 0.1836 μg/g, Cd 0.02795 μg/g, Ni 0.0854 μg/g, Cr 0.0474 μg/g, and Pb 0.2612 μg/g.

Table 2 Atomic absorption spectrophotometer conditions
Full size table

Assessment of contamination

There are several indices that may be used to assess the level of contamination of heavy metals in road dust. For this study, the geo-accumulation index (I-geo) and the contamination factor (CF) were selected to evaluate the contamination level of Zn, Cu, Cd, Ni, Cr, and Pb in the roadside dust, based on their common usage in similar work (e.g. Addo et al. 2012; Awadh 2013). The I-geo assesses the level of contamination by comparing the concentration of heavy metals obtained to background levels which are bottom sediments in this case. The I-geo index was calculated using the following equation:

$$ {\mathrm{I}}_{\mathrm{geo}}={\log}_2\left(\frac{\mathrm{Cn}}{1.5\mathrm{Bn}}\right) $$
(1)

In the I-geo index, Cn is the measured concentration of the heavy metal in the road dust sample and Bn is the geochemical background concentration of the heavy metal (Robertson et al. 2003; Taylor 1964). The constant 1.5 in Eq. 1 minimises the effect of possible variations in the background values which may be attributed to lithologic variations in the sediments. The values of this index vary from sub-zero to more than 5. The highest grade (6) reflects a 100-fold enrichment and (0) reflects the background concentration. 0–1 indicates unpolluted to moderate pollution, 1–2 indicates moderate pollution, 2–3 indicates moderate to strong pollution, 3–4 indicates strong pollution, and 4–5 indicates strong to extreme pollution, while values > 5 indicates extreme pollution.

For comparison with the I-geo index, the CF was also calculated for all samples. The CF was calculated as:

$$ \mathrm{CF}=\frac{{\mathrm{Cm}}_{\mathrm{sample}}}{{\mathrm{Cm}}_{\mathrm{background}}} $$
(2)

CF is the contamination factor, Cm sample is the concentration of the metal pollutant in the sample, and Cm background is the metal concentration in the background sample. The CF reflects the metal enrichment in the sediment; a CF < 1 indicates low contamination, 1 ≤ CF < 3 indicates moderate contamination, 3 ≤ CF ≤ 6 indicates high contamination, and CF > 6 indicates very high contamination.

PM2.5 sampling procedure

Sampling of PM2.5 was carried out at the five sampling sites along Thika superhighway during the months of March and April 2017 (dry season). The background and roadway monitoring of PM2.5 were undertaken using a portable air sampling system, BGI (Kinney et al. 2011). Samples were collected on pre-weighed Teflon filters for 8 h per sample at a flow rate of 4 L per minute, with air being drawn by a calibrated BGI personal sampling pump. At each site, sampling was conducted between 06:30 and 14:30 h local time, for five non-consecutive weekdays. During the sampling period, humidity ranged between 57 to 69%, temperature 21 to 23 °C, and wind was between 8 and 14 km/h.

Results and discussion

Road dust chemistry

The heavy metal (Zn, Cu, Cd, Cr, Ni, and Pb) concentrations in road dust are listed in Table 3, and the spatial distribution of these metals in road dust is illustrated in Fig. 2.

Table 3 Mean heavy metal concentrations in the roadside dust (n = 3) with standard deviations given in brackets
Full size table
Fig. 2
figure 2

Concentration of heavy metals (Zn, Cu, Cd, Ni, Cr, and Pb) along Thika superhighway in μg/g. The background concentration of each metal has been subtracted, implying that the concentration level is the enrichment of that metal at the site

Full size image

The concentration of most of the metals (e.g. Zn, Cu, and Pb) are higher in sites closer to the urban areas of Nairobi City and Thika town (Fig. 2), which is attributable to the greater volume of vehicular traffic near the highly populated conurbations. All these metals were suspected to originate from vehicular emissions, since there was the absence of any close industries near the sampling sites (Victoria et al. 2014; Nabulo et al. 2006). In the case of Allsops (AL) sampling site, it had the highest Zn concentration with an average of 262.2 μg/g which is comparable to Dhaka and Amman studies (Ahmed and Ishiga 2006; Al-Khashman 2004), whereas Juja (JJ) had lower concentration with an average of 154.2 μg/g. The background site had lowest Zn concentrations (94.9 μg/g) when compared to the other roadside sites. AL site had the highest Cu concentration, with an average of 101.4 μg/g, and JJ was lower with an average of 54.1 μg/g, compared to 39.2 μg/g at the background site. These results are above those obtained in Ketu-South, Ghana (60.53 μg/g) showing that the sites in this study are highly polluted (Addo et al. 2012). The site with the highest concentration of Cd was AL with a concentration of 277.0 μg/g, while the lowest concentration of 11.9 μg/g was measured at Commercial Street Thika (THK). For comparison, the background site had a concentration of 8.7 μg/g. In the case of Ni, JJ had the highest mean concentration of 23.5 μg/g while THK had the lowest concentration of 14.3 μg/g. The background Ni concentration was slightly higher than THK at 16.4 μg/g. In terms of Cr concentration, THK was highest with an average concentration of 29.1 μg/g; this was close to the concentration obtained in Luanda, Angola which was reported to be 26 μg/g (Ordόñez et al. 2003). The lowest concentration was at Ruiru (RR) with an average of 13.3 μg/g, which was slightly above the background site concentration (13.0 μg/g). The site with the highest Pb concentration was THK with an average of 72.5 μg/g and the lowest was RR (average of 39.9 μg/g), while the background site had an average concentration of 20.5 μg/g. This is comparable to Kampala, Uganda whose Pb concentration was  45.3 μg/g (Nabulo et al. 2006).

The I-geo shows that most of the road sites are moderately polluted by Zn and Pb, with I-geo scores ranging from 0 to 2, except for Cd which had all I-geo values greater than 5, indicating extreme pollution in the road dust. In the case of Ni, Cr, and Cu, nominal or negative values for most sites indicated that the sites were not enriched with those metals when compared to the background site. The CFs for Zn, Cu, Cd, Ni, Cr, and Pb were calculated for the study area (Table 4). In most of the sites, there was moderate contamination of all metals with CF values ranging between 1 and 3. Cd and Pb had some CF values which were over 3, indicating the road dust samples were more contaminated with these metals, for example, Al site had a CF of 3.1 for Cd, while THK site had CF value of 3.9 for Pb. The lowest CF of all the metals was at JJ site whose CF was 0.6 for Cr, which indicated low contamination. The source of these pollutants could be attributed to exhaust and non-exhaust vehicular emission, since there are no industries close to Thika superhighway except at RR. No anomaly was observed in the RR results that could be attributable to industrial emissions (Xia et al. 2011).

Table 4 Enrichment factors for heavy metals in Thika superhighway road dust using the geo-accumulation index (I-geo) and CF analysis
Full size table

Pearson’s correlation coefficient (r) identified that there was a positive correlation between most of the measured heavy metals in the road dust (Table 5), which indicated a common source for the metals (Victoria et al. 2014). Elemental pairs Cu/Zn (r = 0.986), Cu/Cd (r = 0.908), and Zn/Cd (r = 0.830) had r values above 0.5 indicating strong correlation with each other. On the other hand, pairs such as Cu/Cr (r = 0.291), Cd/Pb (r = 0.107), and Pb/Ni (r = 0.356) showed moderate relationships. Elemental pairs Cu/Ni (r = − 0.054), Zn/Ni (r = − 0.061), Ni/Cd (r = − 0.258), and Cd/Cr (r = − 0.111) indicated inverse relationship. The results showing positive correlation indicate that these metals originated from a common anthropogenic source most probably vehicular emission (Abah et al. 2014; Addo et al. 2012). This was expected as Zn, Cu, Ni, Cr, and Cd are commonly produced from brake wear and tyre wear (Shinggu et al. 2010). Similarly, Cu and Cr have been reported to originate from exhaust emissions (Yu et al. 2003).

Table 5 The correlation coefficient, r, between concentrations of heavy metal along Thika superhighway (Kenya)
Full size table

Particulate matter (PM2.5)

PM2.5 concentrations generally increased towards Nairobi town (Fig. 3), with the highest mean concentration being at JJ with 509 μg/m3. This was attributable to dust storms (Al-Khashman 2004; Amato et al. 2009; Lindén et al. 2012), which were experienced on one sampling day at JJ which skewed the site average (Fig. 4), and therefore, the mean PM2.5 concentration was recalculated for the four remaining sampling days giving 36.24 μg/m3 (Fig. 3). Most of the study sites had high PM2.5 concentrations during Monday, Wednesday, and on Friday (Fig. 4), which could be attributed to the high human-related activities along the highway as people are coming back to the city from their rural homes and as well going back to their rural homes on Mondays and Fridays, respectively (Kinney et al. 2011).

Fig. 3
figure 3

Mean PM2.5 concentrations at five sites along Thika superhighway. All bar graphs show the average PM2.5 of five sampling days, except JJ where one sampling day has been omitted from the calculation due to a dust storm which caused an 8-h average of 2365 μg/m3 on that day

Full size image
Fig. 4
figure 4

Eight-hour PM2.5 concentrations at five sites along Thika superhighway from Monday to Friday

Full size image

PM2.5 concentrations were higher than those observed in many European and American studies, but was similar to studies undertaken in other African countries; Accra, Ghana (Aboh et al. 2009), Ouagadougou, Burkina Faso (Boman et al. 2009), and Cairo, Egypt (Boman et al. 2013). All the sites in this study had lower PM2.5 concentrations when compared to a study completed in February 2006 between Nairobi and Ruiru, Kenya, which was conducted by Van Vliet and Kinney 2007 who reported PM2.5 concentrations of 414 μg/m3. This difference may be due to different sampling protocols and sampling periods employed for the two studies. However, the average PM2.5 concentration for Thika town measured in this study (69 μg/m3) was higher than observed Gaita et al.’s (2014) study in Nairobi whose average concentration was 21 μg/m3, but it was lower when compared to the results (98.1 μg/m3) reported by Kinney et al. (2011) in a side walk within Nairobi central business district. Even though the results obtained cannot be compared with WHO daily guidelines since sampling was done for only 8 h, all the roadsidesampling sites had PM2.5 levels higher than daily WHO guideline of 25 μg/m3.

Health implications

Exposure to heavy metals in roadside dust occurs by skin contact, inhalation, and/or ingestion. The effects of heavy metals in road dust include respiratory system disorders, nervous system interruptions, endocrine system malfunction, immune system suppression, and the risk of cancer in later life (Atiemo et al. 2011). The high levels of heavy metals as evident in this study and particularly Pb and Cr which are known to be carcinogenic are of concern in relation to human health, chiefly to children, the vulnerable old and pregnant women living close to and/or using busy roads (Du et al. 2013; Atiemo et al. 2011).

A great number of scientific studies have linked exposure to particle pollution with a variety of problems including premature death in people with heart or lung disease, non-fatal heart attacks, irregular heartbeat, aggravated asthma, and decreased lung function (USEPA 2014). In addition, increased respiratory symptoms such as irritation of the airways, coughing, or difficulty in breathing have also been linked to PM exposure (USEPA 2015). For those people who spend a significant portion of their day on (or adjacent to) the road, e.g. residential homes, street vendors, traffic police, touts, and public vehicle drivers, the health consequences of PM2.5 exposures are even greater (Ngo et al 2015).

Study limitations

The study only covered a short period in the dry season and a more comprehensive study needs to be executed along the highway and in adjacent residential areas near the highway. It is important to acknowledge that the roadside dust chemistry results only represent part of the dry season.

Similarly, PM2.5 results reported in this study are just representative of a portion of the dry season. Regardless of these, the results show that Thika town as well as Thika superhighway is polluted by PM2.5 and poses a health risk to people working in the town and along the highway respectively, who are exposed for long hours and long periods of time (USEPA 2015; USEPA 2014; WHO 2014, 2015).

The sampling time for PM2.5 was done for only 8 h a day and not 24 h at each location and so the results of are not fully comparable to WHO guideline.

Conclusion

This study has highlighted the influence of vehicular traffic on road dust chemistry and PM2.5 concentrations along Thika superhighway in Kenya. All the heavy metals analysed (Cu, Zn, Ni, Cd, Ni, and Pb) in the road dust samples were above the background concentrations at the sampled sites (except Nickel at one sampling site), highlighting the importance of anthropogenic emissions from vehicular traffic. Since roadside dust may be resuspended and can travelled upto 50 meters, these metals especially Cd, Pb and Cr which had higher I-geo and CF values, pose a significant health risk to people along the highway who may inhale the resuspended polluted dust (Adnan et al 2002). Though not directly comparable, as an indication of air pollution levels, all the sites had PM2.5 8-h concentrations above the 24-h WHO guideline value of 25 μg/m3, meaning that some individuals may experience adverse health impacts which include respiratory and cardiovascular morbidity plus mortality from cardiovascular, respiratory diseases, and from lung cancer (WHO 2013). Those that use the road frequently, including traffic police, street vendors, and local residents along the highway may experience the greatest health impacts due to higher exposures. While this study provides some evidence of the problem, further work needs to include source apportionment studies to quantify the contribution of vehicular traffic to road dust and PM2.5 at the sites in Kenya. Further, while this study has provided initial evidence of PM2.5 concentrations above guideline values, long-term PM2.5 measurements at key sites in Kenya (e.g. important roadways such as Thika superhighway) are needed to more fully understand spatial and temporal variations in particulate air pollution in SSA countries.