Assessment of Pollution Levels of Suspended Particulate Matter on an Hourly and a Daily Time Scale in West African Cities: Case Study of Ouagadougou (Burkina Faso)

In Western countries, research works on air quality have reinforced in recent years because of the links between the level of particulate pollution in numerous cities and the appearing of various health disorders including car-dio-respiratory pathologies, acute bronchopneumonia, lung cancer, etc. In sub-Saharan Africa countries, particularly Burkina Faso, there is very few similar research. In the present work, the pollution levels of airborne particle in the city of Ouagadougou have been assessed through two campaigns of in situ measurements of suspended particulate matter concentrations. These measurements which have concerned PM 1 , PM 2.5 and PM 10 were performed using a portable device (AEROCET531S) at nine sites in 2018 and at ten sites in 2019. These sites are located on roadside, administrative services, second-ary education establishments and outlying districts. The results show that: 1) the PM 1 concentrations values presented no significant variation between days, seasons or sampling sites; 2) the 24-hour PM 2.5 concentrations often ex-ceeding WHO recommended concentrations and, 3) the 24-hour


Introduction
Air pollution is a serious problem that affects the life of billions of people every year (Louati, Son, and Chabchoub, 2018); (Son and Louati, 2016). Africa has been singled out by UN HABITAT as the fastest urbanizing continent in the world (UN-HABITAT, 2010). This fact is accompanied by pollutions, particularly those of air by particulate matter (PM) emission that can cause multiple adverse long-term as well short-term effects on the human wellbeing such as increased health problems (Li et al., 2018) (Chen et al., 2019) (Cassee et al., 2013) (Beltrando, 2014) (Chang, Peng, and Dominici, 2011). In the same vein, and according to the World Health Organization (WHO), more than 25% of deaths around the world may directly be linked to pollution.
The composition and size distribution of particles depend on their formation processes, including their source that has been explored in numerous studies (Tsai et al., 2015).
According to Chatoutsidou and Lazaridis (Chatoutsidou and Lazaridis, 2019), PM may be classified into two groups based on their sources: 1) naturally originated PM and 2) anthropogenic originated PM. The first group includes PM that are emitted by natural sources such as sea spray, volcanoes, forests, and deserts.
On the other hand, common anthropogenic sources are power plants, industries, aviation, vehicles, re-suspension, processes that utilize combustion (the use of biomass as domestic energy, common waste burning practices in residential areas). Previous studies showed that road dust emissions can increase PM 10 by 21% -35% at traffic stations, 17% -34% at urban administrative sites, 17% -22% at industrial sites and 9% -22% at rural sites (Amato et al., 2016).
Meteorological parameters such as temperature, humidity, wind speed and direction play a crucial role in air pollution mitigating (Radaideh, 2017) (Kliengchuay et al., 2018) (Janae et al., 2014). In normal weather conditions, the temperature decreases with altitude so that the pollutants emitted on the ground rise and disperse. This physical phenomenon fades as soon as there is a temperature inversion that favors an accumulation of pollutants in the air, especially in urban environment because at the ceiling of inversion, the pollutants will not be able to disperse any more (Sarr et al., 2018). This would result in containment of pollutants and an increase in concentration at the beginning of the night (Petäjä et al., 2016). Lindén and coauthors (Lindén, Thorsson, and Boman, 2012)  Ouagadougou's climate. It was pointed out that PM 10 levels were substantially higher during unstable weather conditions compared to moderately stable atmospheric conditions across selected locations with various land cover, land use, and traffic density. A similar relationship in the morning is discussed by Etyemezian and coauthors (Etyemezian et al., 2005) which links the largest peak of air pollution to Addis Ababa to a higher atmospheric stability in the morning caused by temperature inversions at the surface during the night. The different intra-urban trends in PM concentrations between day and night can be explained by the difference between the sources of PM. Indeed, there is probably a greater influence of traffic dust suspension on paved and unpaved roads, the exhaust emissions in the morning to that adds the effect of using biomass as a source of energy as we move forward in the day. In addition to these sources, we can list the contribution of the dry season by the Harmattan of Sahara dust and local dust. More stable night conditions favor a mixture of suspended dust with particles generated by combustion and circulation, resulting in more uniform levels of PM 2.5 in the evening.
Meteorological parameters of Ouagadougou are described in section 2.1. In general, wind speeds are low and stable and this is favorable for a stagnation of pollutants in the air, especially after 6:00 pm (Eliasson, Jonsson, and Holmer, 2009). However, the main origin of air pollutants and air pollution, is urban such as re-suspension related to traffic on paved and unpaved roads (Boman et al., 2009). Indeed, according to Boman and coauthors (Boman et al., 2009), most of the geological material found in PM 10 is due to dust suspension from roads related to the prevalence of unpaved roads, the use of biomass as domestic energy and waste incineration. Some natural sources contribute also to PM such as the Saharan desert, which is the world's largest source of wind dust (Goudie and Middleton, 2001), the Bodélé's depression in Chad, which contributes to inject an important quantity of dust transported by Harmattan in the atmosphere of West Africa. A study of long-range dust transport shows that West Africa is the region the most affected by dust transported from the Saharan desert but the least studied (De Longueville et al., 2010).
Some studies on air pollution in the city of Ouagadougou show that this pollution is mainly due to PM and hydrocarbons (Eliasson, Jonsson, and Holmer, 2009) (Boman et al., 2009) (Nana et al., 2012)  . The PM concentrations exceed two or three times the recommended concentrations. Eliasson and coauthors (Eliasson, Jonsson, and Holmer, 2009) were assessed the PM 10 concentrations to 578 µg•m −3 in central business district, 1123 µg•m −3 in high standing residential and 1884 µg•m −3 in traditional residential. Nana and coauthors (Nana et al., 2012)    have measured PM 10 concentrations to 162 ± 144 µg•m −3 and 69.0 ± 46.6 µg•m −3 for extreme and moderate pollution situations, respectively, in dry season 2007. These concentrations exceed the 24-hour PM 10 concentrations recommended by the WHO and European Environment Agency (EEA) (50 µg•m −3 ), as well as United States Environmental Protection Agency US EPA (150 µg•m −3 ). However, these concentrations are lower than the 24 hours total suspended particles recommended limit of 200 -300 µg•m −3 by Burkina Faso authorities (Presidence du Faso, 2001). It will be noticed that there are no recommended limits especially for PM 2.5 , PM 10 and other in Burkina Faso.
In this paper, we present an analysis of PM (PM 1 , PM 2.5 and PM 10 ) concentrations in Ouagadougou for measurement campaigns in 2018 (dry and rainy season) and 2019 (rainy season). PM mass concentrations were analyzed by hour, day and location. The characterization of traffic fleet composition was also described. The overall objective of this study was to address the present status of air pollution due to suspended particulate matter in Ouagadougou's city, ten years after the last status. It should be noted that this study covered more measurement sites than any previous study and was conducted over two years.

Description of Study Area and Sampling Points
Ouagadougou the capital of Burkina Faso, located at 12˚22 North, 1˚31 West, 300 m above sea level, is situated in Sahelian region of West Africa. Its population was estimated at 1,700,000 in 2010 and 2,684,052 in 2020 by National Institute of Statistics and Demography. This corresponds to a population increase of 57.9%. Thus, UN-HABITAT (UN-HABITAT, 2010) has renamed it the most dynamic city in the world. As the city is located in a warm semi-arid climate of Sahel, the climate consists of a dry season from October to May and a rainy season from June to September. During the rainy season, the rainfall ranged between 600 -900 mm, while the dry eight-month period generally receives less than 100 mm of rain . Stable night-time atmospheric conditions are common at the beginning of the dry season in Ouagadougou (Lindén and Holmer, 2011), which are favorable to higher pollution levels (Boman et al., 2009). A study of the local wind field by (Lindén and Holmer, 2011) showed that wind speeds are generally very low in Ouagadougou, thus preventing good ventilation of urban air followed by the dispersion of pollutants emitted locally.
During the dry season, the influence of dust carried by Harmattan winds from Saharan desert in the North and North-East affects the entire Sahelian region and creates important seasonal differences in suspended particulate concentrations.
The highest levels of airborne particulate matter are generally observed in February and the lowest levels in August (Prasad, 2011), (Titcombe and Simcik, 2011). According to Ouagadougou Meteorological Office, visibility is generally reduced by almost half during the dry season compared to the rainy season  Five sites (IRSAT, ZI, BV, Kar and UJKZ) were added in 2019 campaign. Sites were selected to cover as much of the city as possible. Table 1 presents the geographic coordinates of the sampling locations in Ouagadougou city.

Road Traffic Characteristic
In 2016, the Ouagadougou town hall has carried out a characterization of road traffic in the city of Ouagadougou by manual counting (Somda, 2018). About 1,003,997 of daily displacements of people that enter and leave downtown. The distribution of vehicles is as follows: 74% motorized two-wheeled vehicles, 18% private vehicles, 7% transit vehicles and 1% heavy trucks.

Measurement Equipment
An analyzer AEROCET 531S has been used. It is a mass profiler and particle counter combined in a small portable battery-powered unit. This analyzer I. Ouarma et al.
measures particulate matter with diameters between 0.3 and 10.0 µm and some others (total suspended particles). Its detection limit is 1.0 µg•m −3 .

Measurement Methods
Measurements of ambient air particulates concentrations were made at each sampling point during at least 12-hours and 48-hours during the measurement campaigns of 2018 and 2019, respectively. The AEROCET-531S was placed at a height of between 1.5 and 2 meters, which corresponds to the average position of the human airways. Each measurement consisted of one-minute concentrations of PM 10 , PM 2.5 and PM 1 and was recorded on a data storage card. Hourly and daily average concentrations reported here are arithmetic means of the respective 1-min readings in µg•m −3 . The measurement relative uncertainties are deduced from the AEROCET measurement accuracy of ±5%. The calculations of the average concentrations and the measurement relative uncertainties were done by Microsoft Excel. The Origin software, version 9 and the QSIS software, version 2.18.28, have been used for graphs and map, respectively.
It should be noted that in 2018 all measures lasted 12 hours. In 2019 at the sites (UJKZ, H6, ZI and B-V), measurements were taken during 48 hours continuously, 72 hours at the sites (BCDG, C3, Kar and G2), 96 hours at the sites (E7, AB and IRSAT). F4-5, RPNU and PK were not sampled in 2019.

Results and Discussion
Limit values for human exposure to particles recommended by the WHO, EEA or US EPA concern the 24-hour concentrations of PM 2.5 and PM 10 . The results that will be presented and discussed will focus on these particles. Results concerning PM 1 were also presented with respect to the fact that they have more significant human health effects. However, there are not recommended limits values.
The results will be discussed and will take into account the grouping made in the material and methods section. Figure 2 presents the hourly concentration profiles obtained in 2019 campaign for PM 1 in Ouagadougou city. These profiles are characterized by two obvious peaks between 3:00 -8:00 am and 5:00 -9:00 pm, respectively. These peaks can be explained on the one hand by the hours of heavy road traffic linked to the activities of the population and on the other hand by industrial activities on in there. It is important to note that large vehicles circulate in the city of Ouagadougou from 10 pm to 5 am (time allowed for these types of track vehicles). There is also the dynamics of the boundary layer (BL), which results in high dilution rates during the day and low dilution rates at night (Lee et al., 2019). The peaks between 5:00 -9:00 pm are more important than those between 5:00 -8:00 am in Journal of Geoscience and Environment Protection some sites. This could be explained by the increase in emissions during the evening, which stretches in a period when the BL is shallow as compared to the morning situation in which the height of the BL increases rapidly after sunrise.

Hourly PM1 Concentrations
The PM 1 concentration was 20 ± 4 and 17 ± 3 µg•m −3 for 2018 dry season and 2018 rainy season, respectively. These values were similar to those obtained by Talbi and coauthors (Talbi, Kerchich, and Kerbachi, 2017) for Alger city. The hourly PM 1 concentrations for educational institutions sites ranged from 5 to 11 µg•m −3 , for peripheral district sites ranged from 8 to 20 µg•m −3 , for industrial sites ranged from 6 to 28 µg•m −3 , for roadside sites ranged from 7 to 25 µg•m −3 and for administrative sites ranged from 7 to 30 µg•m −3 . These results highlight the importance of the activities of these sites except scholar ones on ultrafine particulates.   PM 10 concentration between 8:00 -11:00 am. Another peak, less important, is observed between 6:00 -7:00 pm. These results are explained by the resuspension due to traffic on unpaved roads in these sites and high (respectively low) wind speed during the morning (respectively the evening). Residential dusty sites located near unpaved roads are characterized by important PM 10 emissions. These areas are also characterized by extensive construction and demolition work, resulting in the re-suspension of dust. The profile of industrial sites shows peak around 9:00 -10:00 pm corresponding to the heavy traffic of large trucks at these hours. The great gear causes a great re-suspension of dust. Roadside sites and administrative ones exhibited low PM 10 concentration because they are associated to paved roads that not generate significant dust resuspension. However, a peak is observed around 11:00 am (respectively 4:00 pm) for roadside sites (respectively administrative sites) and related to worker's movements for lunch (respectively for worker's movement to home).

Seasonal and Spatial Pollution Variability
Tables 3-6 present results of descriptive statistics of hourly (or daily) of PM pollution levels for sampling sites. The percentage of hours in which recommended limits by US EPA, EEA, WHO and Burkina Faso (BFA) were exceeded is also presented. The arithmetic means of hourly concentrations of PM are higher in dry season than rainy season (see lines 1 and 2 of Table 3 and Table 4). Indeed, the data in      Table 3 and Table 4. As previously, this result shows a spatial variability of PM pollution level.
Concerning the daily concentration, where only measurements realized during the rainy season in 2019 (Table 5 and Table 6), the arithmetic means of PM 2.5 (respectively PM 10 ) concentrations varied from 5.2 to 35.0 µg•m −3 (respectively 15.9 to 318.1 µg•m −3 ) overall sites. For all sites, the average percentage of 24-hour concentrations higher than the WHO recommended limit of PM 2.5 (respectively PM 10 ) is about 32% (respectively 69%). This means that the PM 10 pollution in Ouagadougou is more acute than the PM 2.5 pollution. Concerning the PM 10 , 69% means that each people of Ouagadougou is each day exposure to 16 hours of pollution level higher than the WHO recommended limit.
Based on Table 2 This may suggest that they come mainly from traffic. Indeed, the activity of this source may not change significantly from a season to another. For PM 2.5 and PM 10 , their concentrations are subject to different degrees of variability. This would suggest attenuation, extinguishment, or reduction of the contribution of some sources of these sizes of particles, and that these two categories of particulate matter would be added by all other sources in addition to those cited above.
It has been observed the effect of the season on particulate matter concentrations over the city of Ouagadougou. Particulate matter concentrations are attenuated by rainfall by leaching air from a large part of these particles especially coarse.
During the dry season, the influence of dust carried by Harmattan winds from the Sahara Desert in the North and North-East affects the entire Sahelian region and creates important seasonal differences in suspended particulate concentrations.
The results also indicate that the PM concentrations vary depending on the type of site and from the rainy to the dry season. It appears that PM pollution levels are higher in the dry season than in the rainy season. These results show the seasonal and spatial variability of PM pollution levels. There are several reasons for these results. On the one hand, the road network on the outskirts of the city (unpaved roads), the mode of transport of the inhabitants on the outskirts (individual transport) and the lack of rain that enhances the activity of other sources including the re-suspension of local dust by the wind. It is important to note that these pollution levels change with particle size. The amplitudes of concentrations are not the same but the fine particles whatever are their amplitudes have the greatest impact on human health.
The highest concentrations were observed in the dry season. Apart from the site of the E7 which is at the limit of the daily standard all other sampling sites exceed the WHO standard for PM 2.5 by at least a coefficient of two. This may be explained by the fact that over this site there are many trees, which particles are deposited by impact. The high values observed at the RPNU can also be attri- be attributed to winds that precede the rain by suspending important quantity of mineral dust. Table 7 reports the PM concentrations of Sub-Saharan cities, from Bamako, Mali (Garrison et al., 2014), Accra, Ghana (Dotse et al., 2012), Dakar, Senegal (Demay, 2011) and Ouagadougou, Burkina Faso   (Boman et al., 2009). PM concentrations vary between cities, even for those with similar climate and precipitation levels. It is well-known that PM emissions are strongly and negatively correlated with the Harmattan winds from the Sahara Desert and unpaved roads. As an illustration, Garrison and coauthors (Garrison et al., 2014) obtained for one day measurement, a 24-hour concentrations of 43 and 210 μg•m −3 for PM 2.5 and PM 10 , respectively, in Bamako where the Harmattan winds are similar to those of Ouagadougou. Dotse and coauthors (Dotse et al., 2012) measured a 24-hour concentrations of 23.3 and 96.6 μg•m −3 for PM 2.5 and PM 10 , respectively, in Accra where the Harmattan winds are lower than those of Ouagadougou. PM 2.5 pollution levels obtained in 2018 for the city of Ouagadougou (57.9 ± 10.5 μg•m −3 ) are close to those for Bamako (43 ± 21 μg•m −3 ). This seems These values are in the same order than those generally obtained on air pollution in West African cities. The reasons are also more or less identical, namely the high density of the population, the evolution of the vehicle fleet and the major part of the road network which is not subject to a limit.