Introduction

According to the State of the Global Air 2019 report, in Kenya alone, air pollution is the eighth leading risk factor for premature death, accounting for nearly 19,000 deaths (5,000 due to ambient air pollution and the remaining due to indoor pollution) in 2017 (State of Global Air 2019). The report goes on to say that the entire Kenyan population lives in areas with PM2.5 (fine particles with a diameter less than 2.5 μm) greater than 10 μg/m3, which is the World Health Organization (WHO) Air Quality Guidelines for healthy air.

In 2014, the Minister of the Environment enacted the Environmental Management and Coordination (Air Quality) Regulations (EMCA) (2014) to control air pollution. The EMCA (2014) specifies national ambient air quality standards for residential and industrial areas and occupational air quality limits to tackle air pollution. Kenya’s air quality governance framework is structured around ensuring the standards are met. However, the deployment of air quality monitors and the creation of publicly available air quality data to enforce the standards have been slow.

Specifically, the Kenya Meteorological Department (KMD) owns and runs a mobile monitoring van that has high-quality instruments to measure a wide variety of pollutants. None of the original data are public, so the extent of monitoring is unknown. KMD has published some academic papers using data collected mostly in Nairobi by this van. The findings from these articles will be discussed later. KMD also operates a monitoring station on Mount Kenya, which is part of the World Meteorological Organization’s (WMO) Global Atmospheric Watch network. The station has instruments for measuring particulate matter/dust and surface ozone, among other air pollutants, and its data is supposed to be public. However, only surface ozone data is currently being reported at this site. KMD has for several years also been collecting and publishing data on surface ozone and vertical ozone profiles in partnership with MeteoSwiss, the Swiss Office of Meteorology and Climatology, at their Nairobi headquarters. Surface ozone is very harmful to human health and such data is exceedingly useful for developing pollution management plans. Finally, Kenya’s National Environment Management Authority (NEMA) requires industrial facilities to contract a set of authorized laboratories with high-quality equipment to report their stack emissions and to make provisions for continuous monitoring in accordance with the 2014 Air Quality Regulations. However, none of this data has been made public. Therefore, little official air quality data exists for Kenya (deSouza 2018).

In countries such as the United States, a relatively dense official monitoring network has provided the infrastructure to track progress on air quality improvements and to stretch the use of the network beyond enforcing the standards—to identify sources and trends of pollution (Lee et al. 2007; Zhang et al. 2014; among others), examine health effects (Dominici et al. 2006; Samet et al. 2000; among others), and evaluate the effectiveness of different policy interventions (Sullivan et al. 2018; among others). Without access to data from such a single, robust network in Kenya, there have been competing narratives about the state of air pollution in the country. A report produced by a renewable energy firm The Eco Experts, widely circulated in the press Kisambe 2017), identified Kenya as the least toxic country in the world accounting for air pollution, energy consumption, and renewable energy, using data from the International Energy Agency and WHO, and has been widely cited in the press. Other studies indicate that the city of Nairobi has severe air quality issues.

This article reviews scientific studies that report air pollution concentration levels of widespread ambient pollutants, PM2.5 (fine particles with diameters less than 2.5 μm), PM10 (coarser particles with diameters up to 10 μm), nitrogen dioxide (NOx), sulfur dioxide (SO2), and tropospheric ozone (O3), for which standards exist in Kenya. Most such studies were conducted in the city of Nairobi. Most were limited to short-duration measurements with different monitoring devices in different locations. However, taken as a whole, there is remarkable consistency in the major sources of ambient air pollution identified. This article also uses the published literature to review which pollutant concentrations exceed the current standards in Nairobi and where these exceedances occur. This information can thus provide policymakers with the information needed to take steps to reduce air pollution in Nairobi.

Methodology

Studies that report widespread ambient pollutant concentrations in Kenya were identified by using the search phrase: “Nairobi OR Kenya” and “air pollution OR air quality OR particulat* OR nitrog* OR ozone,” in the PubMed, Web of Science, and SCOPUS databases on June 01, 2020.

Although most air pollution studies were conducted since the early 2000s, there are some studies that go back as far as the early 1990s. Some of the earlier studies reference results from ground-based pollution measurement campaigns that had been carried out in the early 1980s. The results from these early campaigns were published in student dissertations and could not be found in the peer-reviewed literature. However, such measurements are critical to providing an understanding of how air pollution in Nairobi has changed over time and are thus included in this review. In order to document studies published in dissertations, local journals, and conference proceedings, searches of the aforementioned databases were supplemented with thesis searches, Google searches, and a careful perusal of references in the early studies to identify gaps that the author may have missed. The inclusion of these measurements and the focus on multiple widespread pollutants distinguish this review from others conducted previously (Petkova et al. 2013; Nairobi Air Quality Action Plan 2019–2023).

While many studies have been conducted on household air pollution from cookstoves in Kenya, the current review is limited to studies of outdoor pollutant concentrations. I also excluded studies that reported concentrations of pollutants other than criteria pollutants: particulate matter (PM), NOx, SO2, and O3, such as volatile organic compounds (VOCs), polycyclic aromatic hydrocarbons (PAHs), and H2S. Section S1 in Supplementary Information contains a short note about these studies. I choose to limit this review to outdoor criteria pollutants because as mentioned in the Introduction, Kenya has standards for certain criteria or widespread outdoor air pollutants and there is a growing understanding of the large global burden of disease caused by criteria pollutants like PM2.5 and O3 (Anenberg et al. 2010).

In addition, multiple studies analyze data from the station deployed on Mount Kenya as part of the World Meteorological Organization Global Atmospheric Watch (GAW) program and operated by the Kenya Meteorological Department. This station is situated far from local pollution sources to measure regional pollution/background pollution levels. The current review highlights work that reports pollutant concentrations from this GAW station. However, I do not include studies that only discuss the setup of this monitoring station (e.g., Henne et al. 2008a), or the comparison of the data from the GAW station on Mount Kenya with GAW stations in other regions of the world (e.g., Klausen et al. 2009).

I also excluded studies that relied on air pollution models rather than ambient measurements. I have retained studies that report concentrations of components of PM—even if they don’t report PM2.5 concentrations, as they provide important information about PM sources in Kenya. Such studies could thus provide important insights to policymakers about how to lower PM concentrations in Nairobi most effectively.

Finally, as such a review relies on studies already published, it does not capture burgeoning citizen science efforts to measure air pollution in Nairobi. Of these efforts, the work by Code for Africa (data available at: https://open.africa/dataset/sensorsafrica-airquality-archive-nairobi, Last Accessed June 9, 2020) is notable and has been widely shared on social media.

Results and discussion

Overall, a total of 33 studies matched the review criteria and are summarized in Table S1 in Supplementary Information, which provides information about the pollutant measured, the kind of instrument, time period and duration of the study, as well as key results.

The timeframes of the studies vary, with most measurement campaigns running for a few weeks to several years. Different instruments, with different detection limits and levels of precision, were used among the studies, making direct comparisons across studies difficult. Four studies (Shilenje and Ongoma 2014; Shilenje et al. 2015, 2016; Apondo et al., 2018) were conducted with funding from the Kenyan Meteorological Department and are indicated in bold in Table S1.

All but nine of the studies focus on measurements in Nairobi alone. Of the nine, four (Gatari et al. 2001, 2005, 2009b; Henne et al. 2008b) report measurement results in Nanyuki only. The GAW site is on Mt. Kenya in Nanyuki. Two report concentrations in Nanyuki in addition to Nairobi (Gatari and Boman 2003; Pope et al. 2018). One study reported pollutant concentration in the Athi River township (Shilenje et al. 2015). One study reported pollutant concentrations in Nakuru in addition to Nairobi (Odhiambo et al. 2010), and one study was conducted in Kiirua in Meru county in rural Kenya (Teather et al. 2015).

Twenty studies report PM1, PM2.5, or TSP concentrations (Ngang’a and Ngugi 1986; Karue et al. 1992; Gatebe et al. 2005; van Vliet and Kinney 2007; Gatari et al. 2009a, b, 2019; Odhiambo et al. 2010; Kinney et al. 2011; Gaita et al. 2014, 2016; Ngo et al. 2015; Shilenje et al. 2015, 2016; deSouza et al. 2017; Mukaria et al. 2017; Maina et al. 2018; Pope et al. 2018; Egondi et al. 2018; West et al. 2020). One study reported  PM concentrations for non-standard fine and coarser particles at two different heights in different locations in Nairobi (Gitari et al. 2010). Another study included in this review also reports the number and kind of bacteria in aerosol samples in Nairobi (Yin et al. 2018). As each of the 20 studies gives pollutant concentrations using different time-averaging, presenting a coherent summary is difficult, especially as all studies except for deSouza et al. (2017)Footnote 1 did not make the raw data publicly available.

Figure 1 displays estimates from sixteen studies of PM1, PM2.5, PM10, and total suspended matter (TSP) in units of μg/m3 averaged over the study period. Average PM concentrations over the study period shown in Fig. 1 were those directly reported in the different studies. The two studies sponsored by KMD provide daily average PM2.5 and PM10 concentrations and are displayed separately in Fig. 2. The WHO daily air quality guideline is displayed, where applicable in both Figs. 1 and 2 for comparison. It can be seen that many studies, including official monitoring efforts by KMD, report harmful concentrations of PM. I display the daily-averaged WHO value, 25 μg/m3, which offers a conservative standard for comparison. The WHO annual-averaged standard for PM2.5 is 10 μg/m3.

Fig. 1
figure 1

(a) PM2.5, (b) PM1, (c) PM10, and (d) TSP in units of μg/m3 for different studies averaged over the study-specific period. Study-specific periods are listed in the legend. The red dotted lines on panels a and c indicate the World Health Organization (WHO) daily guideline. As there is no WHO guideline for TSP, the red dotted line on panel d indicates the Kenyan 2014 daily average standard for residential areas. There is no standard set for PM1

Full size image
Fig. 2
figure 2

Daily-averaged (a) PM2.5 and (b) PM10 reported by scientists at the Kenya Meteorological Department (KMD) using an official mobile air quality monitor owned by KMD. The red dotted lines indicate the WHO PM25 daily-averaged guideline

Full size image

Almost every published study that measures particulate matter (PM) going back as far as 1983 in Nairobi indicates that levels of some air pollutants in certain parts of the city are unsafe vis a vis the World Health Organization (WHO) guidelines. For example, Ngugi (1983) (The same results also found in Ngang’a and Ngugi (1986)—I have thus combined the two studies in a single row in Table S1) reports that estimates of TSP reached levels exceeding 250 μg/m3 in the industrial area, where many poorer residents of Nairobi live. This estimate exceeds Kenya’s 2014 TSP standard for residential areas, which is 200 μg/m3 for 24-h averages and 140 μg/m3 for annual-averaged measurements. Fast-forwarding to 2015, West et al. (2020) find PM2.5 concentrations as high as 300 μg/m3 in Viwandani, a poor neighborhood in the industrial area. Average PM2.5 concentrations over about a month experienced by the residents followed in the study ranged between 22.4 and 39.6 μg/m3, which is almost equal to or higher than the WHO daily PM2.5 guideline of 25 μg/m3. Recent research that uses visibility measurements as a proxy for PM2.5 estimated that PM2.5 concentrations increased by ~ 180% between 1970 and 2018 in Nairobi (Singh et al. 2020).

Many of the studies also indicate likely air pollution sources based on different techniques such as (i) deciphering the diurnal and seasonal measurement trends (deSouza et al. 2017, among others), (ii) extracting source signatures from the aerosol elemental composition (Ngo et al. 2015, Yamamoto et al. (2015) among others), and (iii) examining the correlation between pollutant concentration and activities such as traffic patterns (Odhiambo et al. 2010; Mukaria et al. 2017). Regardless of the technique, all the studies reviewed show a remarkable consistency in the dominant sources identified.

Specifically, all studies demonstrate that vehicular emissions are a highly important pollution source in Nairobi. Roadside pollution concentrations were much higher than in residential/background areas. Levels of air pollution along major roadways can be dangerously high, in almost all cases exceeding the Who daily guideline as shown in Figs. 1 and 2. One study that examined air pollution exposure among bus drivers and mechanics that worked on or close to roads found extremely high exposure rates (Ngo et al. 2015).

Nine of the 33 studies also report black carbon (BC) (Gatari and Boman 2003; van Vliet and Kinney 2007; Gatari et al. 2009a, b, 2019; Gaita et al. 2014; Ngo et al. 2015; Shilenje et al. 2015, 2016): a component of PM produced from the incomplete combustion of fuel (typically produced by older vehicles) that causes many adverse health effects (Jansen et al. 2005; Janssen et al. 2011). A key finding from this review is that BC forms a large percentage of the PM in Nairobi, with levels among the highest reported in the world. Mbandi et al. (2019) provide estimates of the fuel economy of road vehicles in Nairobi. They find that the fuel economy of different vehicles is 2–3 times worse than in the countries from which these vehicles were imported: Japan, Europe, India, and China. Specifically, this article notes that Kenyan passenger vehicles would require a 4-fold improvement to achieve the Japanese fuel economy standard of 5.95 L/100 km.

Multiple studies showed that the informal neighborhoods, especially the ones in the industrial area, suffer from particularly high levels of pollution. This can also be seen in Fig. 1, where the neighborhoods in the industrial area, Korogocho, and Kibera exhibit fine particle levels of several 100 μg/m3. Code for Africa has partnered with journalists, providing them with low-cost air quality monitors so that they can monitor specific factories that residents have long complained about. This work resulted in the temporary closure of an asphalt factory in the neighborhood of Syokimau in Nairobi (Serwanjja 2019). This initiative points to the value of scientists partnering with communities to achieve common goals.

Nearly one-third of the studies also report concentrations of other trace elements such as lead (Pb), bromium (Br), zinc (Zn), and manganese (Mn), which were used to identify pollution sources. Invariably most studies performed over time find that resuspended road dust, the wear and tear of brake linings, biomass burning, and emissions from industrial sources are other major sources of PM in Nairobi. Several studies going as far back as the early 1990s reported much higher levels of Pb in Nairobi. For example, Karue et al. (1992) reported Pb levels in Nairobi that were as much as 5 times higher than in European countries at that time. Gatebe et al. (1996) demonstrated that vehicle emissions were major sources of Pb and Br. Gatari et al. (2005) reported high levels of Pb near open-air vehicle garages. Odhiambo et al. (2010) also found levels of Pb and Mn (an additive to petroleum fuels) at a busy roundabout in Nairobi to be much higher than levels in the United States or Europe. Gatari et al. (2009a) found that Mn levels in Nairobi exceeded the WHO guideline of 150 ng/m3. All the aforementioned studies were conducted before 2006, the year when leaded gasoline was phased out. However, even after 2006, Maina et al. (2018) found high concentrations of Cd, Pb, and Zn at locations along Thika highway, indicating moderate to high contamination. Gaita et al. (2014), from their examination of the ratio of Br/Pb in Nairobi, concluded that tire abrasion, likely made worse due to the heavy use of secondhand tires, was a major source of aerosol in Nairobi. More work is required to investigate the sources of the heavy metal concentrations in aerosols in Nairobi and elsewhere in Kenya.

The studies conducted so far were in a small number of locations within Nairobi and in a few other towns. In order to obtain countrywide estimates of pollution, annual-averaged population-weighted PM2.5 estimates for Kenya derived from satellite data calculated using the method listed in van Donkelaar et al. (2016) (Publicly available from: http://fizz.phys.dal.ca/~atmos/martin/?page_id=140 , Last Accessed Feb 17, 2020) for the period 2000–2016 are displayed in Fig. 3, in conjunction with new vehicle registrations (https://www.ceicdata.com/en/kenya/road-transport-number-of-motor-vehicles-registered/road-transport-no-of-motor-vehicles-registered, last accessed February 17, 2020). This extensive dataset allows us to see the long-term trend of increasing PM2.5 as well as the correlation between satellite-derived PM2.5 and vehicles registered. However, it must be noted that satellite estimates of PM2.5 are error-prone, especially in areas without reference monitors. The dip in PM after 2006 is likely due to the fact that Kenya phased out leaded gasoline in 2006. More work is required to characterize the long-term effects of such policy interventions. The caveat with using this data is that Alvarado et al. (2019) found high errors in satellite-based estimates of daily PM2.5 in cities in low- and middle-income countries. They highlight that satellite-derived data can never be a substitute for reference ground monitors in such cities. However, trends derived from long-term satellite data records are much more robust than instantaneous values.

Fig. 3
figure 3

(a) Population-weighted annual PM2.5 levels in units of μg/m3 in Kenya from 2000 to 2016, derived from satellite data, using the method described in van Donkelaar et al. (2016). Note these results produce significant uncertainties, but the qualitative trends are likely to be robust. (b) Number of new motor vehicles registered/year in Kenya from 2008 to 2017 from CEIC Global Economic Data, Indicators, Charts and Forecasts data

Full size image

In contrast to PM, eight studies report concentrations of different gaseous pollutants (Henne et al. 2008b; Odhiambo et al. 2010; Shilenje and Ongoma 2014; Shilenje et al. 2015, 2016; deSouza et al. 2017; Kimayu et al. 2017; Apondo et al. 2018). Each of these studies conducted so far finds that gaseous concentrations fall within the WHO guidelines. Across the different studies, researchers find that vehicles are a major source of NO and NO2. NO2 is likely an important source of tropospheric O3. The formation of O3 is strongly dependent on solar radiation and temperature. Apondo et al. (2018), however, found that concentrations of CO, likely formed from vehicular emissions, exceeded the WHO guidelines near major urban roads. More studies are needed to better characterize the dynamics of gaseous formation in Nairobi and to better understand the sources.

Each panel in Fig. 4 shows the concentrations of gaseous pollutants for which multiple averages are reported at different locations for the same time-averaged intervals. For example, Henne et al.’s (2008b) study is the only one that reports night-time concentrations of CO; Shilenje and Ongoma (2014) study is one of the few studies that report the annual average concentration of O3 and have thus been excluded from graphical representation.

Fig. 4
figure 4

Daily- and annual-averaged CO, SO2, NO2, and O3 concentrations reported in different studies

Full size image

Conclusions

Existing measurements of air quality in Nairobi indicate that for particulate matter (PM), the WHO guidelines have been exceeded in many parts of the city. Moreover, the guidelines have been violated for many years—going all the way back at least to the early 1980s. There is thus an urgent need for action to improve air quality in Nairobi.

Although the government of Kenya has not made the original data collected from official monitoring systems publicly available, the analysis of such data has been published in scientific journals. In Table S1, I highlight the studies that have been funded by the Kenyan Meteorological Department. These studies confirm that PM levels in Nairobi violate the WHO air quality standards. There is thus a need for transparency in official monitoring efforts and for the data to be made publicly available, for scientific as well as policy development purposes.

Another important finding from this review is that vehicles are a major source of pollution in Nairobi. This result can inspire concrete action by incorporating concerns about air pollution more fully into the environmental impact assessment (EIA) of new transportation-related projects in Nairobi. Air pollution concerns can also lead to a realization of the need for more projects that promote non-motorized transport (NMT), the improvement of public transportation infrastructure, and the need to enforce the ban on importing older vehicles into the country. The Kenyan government has taken first steps to impose restrictions on importing such vehicles (Miriri 2019). It is important that such policies are enforced. Scrutiny of sources of heavy metal concentrations in Nairobi is also required to evaluate the need to speed up moving to a higher fuel standard.

A gap in the existing literature that this review highlights is that most of the studies conducted so far have been in Nairobi. More measurements in other urban areas in Kenya such as the busy Mombasa port need to be conducted. Satellite measurements can be used to help identify hotspots of pollution to guide the placement of more monitors. However, satellite measurements can never be a substitute for high-quality reference monitors, as previous research has found that satellite-derived PM2.5 estimates for many cities in the global South can have larger errors (Alvarado et al. 2019). The errors are lowest in inland locations that are at low altitudes (Alvarado et al. 2019). deSouza et al. (2020) presents a new technique to constrain satellite estimates of PM2.5 using low-cost sensors that have the potential to produce more accurate estimates of PM from satellite data in Nairobi. There is thus a need to expand the ground-based air quality monitoring network in Kenya.

In addition, very little information exists about the trends of gaseous pollutants in Nairobi and elsewhere in Kenya. This is another gap that needs to be filled. Many of the studies in this review were conducted for short periods of time. For example, one of the few studies that measures gaseous pollutants only comprises of measurements for December 25 and 26, 2015 (Shilenje et al. 2016). This finding highlights the need for long-term continuous measurements of air pollutants in Kenya to track air pollution trends in the city. Such a data source can also allow researchers to evaluate the impacts of different policies.

Even within Nairobi, more work is required to characterize the exposure of vulnerable populations in Nairobi, such as those that live in the informal areas, where studies so far have found high levels of pollution (Ngo et al. 2015; deSouza et al. 2017). The contribution of different sources to pollution levels in these areas must be assessed for urgent action to be taken to reduce concentration levels.