Source of polycyclic aromatic hydrocarbons (PAHs) in rainwater and effect on the health of the population: the case of the District of Abidjan in the South of Ivory Coast

Anthropogenic activities represent a real source of pollution for the environment and their multiplication during the last decades generates increased contamination of the environment (Hansen et al. 2003). Polycyclic aromatic hydrocarbons (PAHs), like many pollutants, have been introduced into aquatic ecosystems (Bihanic 2013) by dry and wet depositions of atmospheric compounds. PAHs are chemical elements formed by a set of carbon atoms and hydrogen atoms; the set is organized into at least two benzene rings (Golly 2014). The major sources of these pollutants are the processes of incomplete combustion of organic matter of fossil origin or biomass, such as wood-burning (Lee et al. 1976). The most studied PAHs pose many environmental problems because of their carcinogenic character and are among the emblematic pollutants of the World Health Organization (WHO 2004) and the European Community (OJEC 2004) in relation to water quality. These chemicals are introduced into the body by water consumption or exposure to air pollution (Karyab et al. 2016). Medical and laboratory findings indicate that the exposure of any duration to PAHs causes serious cancer problems (Anderson et al. 2000). Previous studies have revealed that the air in Abidjan is polluted (WHO 2005; Kouassi et al. 2012; Bahino et al. 2018; Gnamien et al. 2020). However, some low-income residents use rainwater that leaches into the atmosphere for their domestic needs. Due to their volatility, these PAHs could be found in these waters even at low concentrations and be the cause of health problems for the Abidjan population. This study, carried out in two communes of the District of Abidjan, Treichville and Cocody, has the main objective of the identification of the potential sources of PAHs measured in the collected rainwater and the evaluation of the health risks to the population. It is the first and provides information on these pollutants (concentrations, sources, and toxic effects) that can contribute to a better approach in the fight against carcinogenic diseases.

The District of Abidjan, with its high population and socio-economic activity, is confronted with the phenomenon of pollution. Indeed, the main sources of pollution, as in many West African capitals, are industrial units, road traffic, domestic fires, and open dumps. For this study, rainwater was collected in the communes of Cocody and Treichville (Figure 1) according to certain criteria. The choice of the Cocody and Treichville sites is justified first by the existence of air pollution measurement sites within the framework of the PASMU project (Air Pollution and Health in Urban Environments of the Ivory Coast), and consequently, the acquisition of scientific data necessary for this study was easier. Then, these two communes have different characteristics. Cocody, where the standard of living is high and its site of measurement is very impacted by the road traffic source; Treichville, with an average standard of living, has a site of measurement where the domestic source is well represented as well as the road traffic source because of the existence of the industrial zone of Koumassi, the lagoon, and the port of Vridi located at a few kilometres.

Rainwater was collected after each rainfall episode during the months of March and June. Ten samples were selected for this study. The rain collector used was developed at the Aerology Laboratory of Toulouse (LA). A rain detector automatically controls the automatic opening of the sensor, which is made up of an inert plastic bag. Each of these sites is equipped with this automatic rain collector, consisting of a cover, a single-use polyethylene sampling bag, and a sensor. These components are mounted on a mast at a height of approximately 1.5 m and placed in an open area to collect rainwater directly. The wet deposits are collected at the event (after each rainfall) by a semi-automatic collector dedicated to the analysis of the chemical composition of precipitation.

The Spearman's correlation coefficients with a significance level taken at p<0.005 were used to establish the relationships between PAHs and the emission sources.

The use of molecular profiles and isomeric ratios between PAHs is used to assess the sources of PAHs in collected rainwater. Numerous molecular indices for the identification of different sources are used in the literature (Colombo et al. 1989; Budzinski et al. 1997; Yunker et al. 2002). The reliability of these ratios is based on the use of pairs of isomers, i.e., compounds with the same molecular weight and with the least differences in their physicochemical characteristics and their evolution in the environment. The Phenanthrene/Anthracene and Benzo(a)Anthracene/Benzo(a)Anthracene+Chrysene pairs are used by some authors to distinguish the pyrolytic source from the petrogenic source (Fauches 2017). Raviendra et al. (2008) use the couples Benzo(a)Pyrene/Benzo(ghi)Perylene, Fluoranthene/(Fluoranthene+Pyrene), and Fluorene/(Fluorene+Pyrene) to distinguish the pyrolytic sources between them. There are many tools and methods developed in the literature for the identification of the origin of these elements in nature. However, sometimes these tools are not always accurate, so the isotopic (δ¹³C) method for the 16 PAHs was developed to confirm the results of molecular ratios.

PAHs are easily measured after extraction from water (Sun et al. 2003; Yan et al. 2006). The compound-specific isotope analysis (CSIA) method mainly targets 16 primary PAHs (EPA method 8310) from two to five rings; other PAHs can also be targeted for measurement upon request. PAHs are dissolved in pentane or methylene chloride, injected at 305 °C (no fractionation, 1 min), and separated on an Agilent DB-5 ms Ultra Inert column (30 m×0.25 mm ID×0.25 μm film thickness) at a constant flow rate of 1.4 ml under the following temperature schedule: 40 °C (1 min hold); 140 °C (10 °C/min); 310 °C (4 °C/min; 15 min hold); 320 °C (10 °C/min; 5 min hold). GC-C-P-IRMS is performed on a Thermo Trace GC 1310 gas chromatography coupled to a Thermo Finnigan MAT 253 isotope meter via a GC IsoLink II combustion interface. For 13C analysis, individual PAHs are converted to CO₂ in a combustion reactor consisting of a NiO tube containing CuO and NiO wires maintained at 1,000 °C. The water is then removed by a Nafion dryer before the gases to be analyzed are transferred to the Isotope-ratio mass spectrometry (IRMS). For 2H analysis, individual PAHs are converted to H₂ in a high temperature thermal conversion reactor consisting of a graphitized Al₂O₃ tube maintained at 1,425 °C.

The QC/QA mixtures are either certified mixtures distributed by Indiana University or are composed of pure n-alkanes that have been calibrated separately by elemental analysis (EA)- and TC/EA-IRMS using certified reference materials (e.g., NBS-22 and IAEA-CH-7) distributed by USGS, National Institute of Standards and Technology (NIST), and International Atomic Energy Agency (IAEA). All of these materials are directly traceable to the primary isotopic reference material for each element (i.e., Vienna Pee Dee Belemnite (VPDB) for δ¹³C). The calibration procedures for PAH CSIA are applied identically for reference materials and samples.

First, the provisional isotopic value of each PAH is obtained by normalization against an isotopically calibrated internal reference compound (e.g., SPEX CertiPrep CLPS-B; e.g., c12:0 FAME or c19:0 FAME). The isotopic values of the individual PAHs are then normalized to the scale of the primary reference materials using IU A7, an external mixture of broad-spectrum n-alkanes with a wide range of calibrated δ¹³C values.

The method for health assessment of PAHs and dermal cancer risk from carcinogenic PAHs using formulas is defined by the U.S. EPA (2007). This formula allowed us to determine the average dermal absorded dose (DAD).

The carcinogenic risk in this study was calculated by the estimation of DAD, incredimental lifetime cancer risk (ILCR), and cancer risk index (R) in the water sample. Equation (2) was used to calculate the DAD, except that AT in the formula has been taken as 25,550 (i.e., 70*365 days) for the two age groups.

This study concerns six carcinogenic PAHs, namely Benzo(a)pyrene, Benzo(b)fluoranthene, Benzo(a)anthracene, Dibenzo(ah)anthracene, Chrysene and Indeno (1.2.3cd) pyrene, and three non-carcinogenic ones, namely Naphthalene, Phenanthrene, and Fluoranthene. The six other PAHs due to the absence of certain data such as Kp were, therefore, not considered in the calculation.

This chapter is devoted to the presentation of the results of the various analyses carried out and to discussions.

The results of the analyses of the different PAHs measured in the collected rainwater and the total PAHs are reported in Tables 2 and 3. The total PAH is the sum of the 15 PAHs analyzed.

The analysis of Table 2 shows that the total PAHs vary from 121 ng/l (rain of March 23 in Cocody) to 4,204 ng/l (rain of 2nd February in Treichville). The average of each PAH taken individually varies from 26.6 ng/l (Ant) to 306.6 ng/l (PHE). PAH concentrations during 2019 range from less than 0.0001 ng/l (DaA) to 1,842 ng/l (PHE). PAH concentrations are measured in the 2020 rainfall and, as summarized in Table 3, are much lower than those in the 2019 rainfall. The accumulations of the 15 PAHs vary from 0.4 ng/l (rain of June 5 in Cocody) to 318.3 ng/l (rain of 1st April in Treichville) and are 100–300 times lower than those of 2019. The average of individual PAHs is between 0.4 ng/l (Ant) and 30.9 ng/l (IP). The concentrations range from less than 0.0001 to 76.6 ng/l. The highest concentration of 76.6 ng/l was recorded for BghiP during the rain of April 1st, 2020 in Treichville. These low averages of Ant in relation to the total weight show its low contribution compared to PHE and IP, which give the highest averages.

The high levels of PHE followed by Na and ACE in the dry period of the first year could be explained firstly by their high volatility and half-life when exposed to ultraviolet radiation. This observation was also made by Fauches (2017) in Paris. Indeed, the author had recorded high values of these PAHs in rainwater. Higher or lower values of heavy PAHs (CHR, BkF, BghiP, and FLA) in the dry season could be justified by the amount of rain collected during these seasons. The dry season was more loaded with atmospheric particles, and the atmosphere was, therefore, more leached with the heavy rainfall recorded during the rainy episodes collected during this period. Motelay-Massi et al. (2003) showed that the composition of the total deposition (dry and wet) shows a majority of volatile PAHs, such as phenanthrene and fluoranthene, compared to heavy PAHs (BkF and BaP) as shown in our study (Figure 2). The same observation was made by Lee et al. (2006) during a sampling campaign in a city in South Korea. These authors showed that the sum of high-molecular-weight PAHs with a low Henry's law constant was less than that of low-molecular-weight PAHs with a high Henry's law constant.

Tables 4 and 5 highlight the evolution of the static variables of the different PAHs studied according to the seasons. During the dry season of 2019, the concentrations vary from less than 0.0001 μg/l (concerning several PAHs) to 1.842 ng/l (PHE). The means and standard deviations range from less than 0.0001 to 759.9 ng/l (PHE) and from less than 0.0001 to 721.4 ng/l (PHE), respectively. The highest concentration is observed for Phenanthrene (1.842 ng/l), while Dibenzo(ah)anthracene could not be analyzed as all values are below the detection limit of the instrument. The highest values of PAH accumulation were observed in the rain collected on February 2 in Treichville and the weakest values in the rain on March 23 in Cocody. In the rainy season, the concentrations vary from less than 0.0001 to 101 ng/l (observed for ACE on February 7 in Treichville). The average PAH over the study period varies from 18.2 ng/l (FLH) to 67.5 ng/l (BaA), and the standard deviation is between 8.9 ng/l (BghiP) and 31 ng/l (Py).

The pie charts below (Figure 2) present the contributions of each PAH individually during the two seasons. In the dry season, PHE remains the dominant element with 41% of total PAHs, while BbF and DaA contribute 0%. In the rainy period, ACE contributes 11% and remains the most abundant. PY and FLA each contribute 3% and are the least abundant.

In 2020, PAH concentrations vary from less than 0.0001 to 76.6 ng/l in dry periods. The highest concentration is observed for BghiP. The average concentration is between 0.4 ng/l (ANT) and 51.7 ng/l (IP), and the standard deviation varies from 0 ng/l (Fl) to 33.3 ng/l (BghiP). During rainy periods, the concentrations vary from less than 0.0001 to 63.7 ng/l. The IP gives the highest concentration (63.7 ng/l). The mean and standard deviation ranged from 0.4 ng/l (ANT) to 38.3 ng/l (BaP) and 0.2 ng/l (ANT) to 19.8 ng/l (PI), respectively.

The pie charts in Figure 3 reveal that during the dry season, IP is the most dominant PAH with an estimated contribution of 23%, while ACE, FL, and ANT contribute 0%. During the rainy season, BaP contributes 18% and remains the majority element. ANT, DaA, ACE, and BghiP remain the lowest with a contribution of 0%.

Variations in PHE levels from one year to another could be justified by the high volatility of this element when environmental conditions are favourable. According to Lampréa (2009), meteorological conditions (rainfall, relative humidity, and temperature) influence PAH concentrations in rainfall. Indeed, heavy rainfall leaches the atmosphere and collects sufficient quantities of PAH. Comparing these parameters during the 2 years, the first year gives a rainfall of 15.6 mm against 11.87 mm in the second year. The relative humidity of the first year is estimated at 89.45%, while the second year gives 65. 86% and the temperature is 26.68 °C in the first year and 27.15 °C in the second year. The decrease of the PHE could also be explained by a loss of volatilization. Indeed, the vapour pressure of PHE being very high, it tends to evaporate more easily than another compound when the temperature increases (LCSQA 2000).

PAHs being the most part carcinogenic and mutagenic, the identification of their sources and the highlighting of their effects (actions) require the study of the most abundant molecules in water. Indeed, PAHs with two or three benzene nuclei are called light PAHs and those whose number of nuclei is between 4 and 6 are qualified as heavy. Figure 4 shows the evolution of PAHs according to the sites and the number of nuclei during the first year (2019). At Treichville (TRE), the proportions of light PAHs are more important than the heaviest ones, except for the rain on June 2, 2019. At Cocody, the proportions of light and heavy PAHs are roughly equal overall. During the dry season, the proportion of light PAH varies from 69 to 76%, while that of heavy PAH varies from 24 to 33% in Treichville. In Cocody (COC), the proportion of light PAH varies from 38 to 55%, while that of heavy PAH varies from 45 to 62%. In the rainy season, the values of light PAHs are generally higher than the heavier ones. The proportions are between 40 and 90% for the light ones. The proportions of heavy vehicles vary from 10 to 60% in Treichville, while in Cocody, the proportions fluctuate between 50 and 60% for light vehicles and 40 and 50% for heavy vehicles.

Figure 5 presents the evolution of PAH during the second year (2020). We note a different evolution from that of the first year, and the lightest ones present rather weak values during the two seasons, except for the rain on June 5 in Cocody. In the dry season, the proportions in light PAH vary from 2 to 5%, while the PAH with strong molecular weight presents proportions included between 95 and 98%. In the wet season, the proportions of light PAH vary from the maximum percentage of 3 to 100% recorded on June 5 in Cocody. The heaviest ones give proportions ranging between 0 and 97% and are the most abundant. This variation would be attributed to PAH emissions of pyrolytic origin and to meteorological conditions. Indeed, the more the temperature decreases (case of wet periods), the more PAHs will tend to be present in the particulate phase (Botta et al. 2014). Also, the work of Létinois et al. (2012) shows that the temporal evolution of concentrations does not follow a regular seasonal cycle as observed in our study in view of the 2 years of collection. These authors have shown that concentration peaks can be observed during the winter season (December or January) or in summer. Moreover, Motelay-Massei et al. (2003) show that in urban areas, sources are much richer in particulate PAHs that are leached into the atmosphere. PAHs with two and three rings (light) are mostly associated with the gas phase, while those with four rings (heavy) and more are mainly associated with the particulate phase (Foan 2012).

The lightest PAHs are, therefore, more present in environments where domestic activities are dominant. The predominance of detri-aromatic PAHs (PAHs with two and three aromatic rings) relative to PAHs with higher molecular weights would be a sign of contamination of petrogenic origin (accidental spillage or not of oil and petroleum). This predominance of these light PAHs would reflect an origin linked to emissions from heavy-duty vehicles using diesel as fuel, whereas light vehicles are the main sources of BaP and D(ah)A which are heavy PAHs (Miguel et al. 1998). This source has been observed in two rainfalls of Treichville with a predominance of Phenanthrene and Acenaphthene. Conversely, the environments mainly marked by pyrolytic pollution (combustion of organic matter like wood, coal, and oil) present a predominance of PAHs with high molecular weight, which are formed under pyrolytic controls. By comparing the PAH concentrations according to the two collection years, the second year (2020) shows low values of light PAH. This would be justified, on the one hand, by the fact that precipitation drains dust as well as pollutants from the atmosphere to the soil with greater efficiency for large particles (Degremont & Cachot 2009). The first rainfall leaches high-molecular-weight PAHs (greater than four rings) associated with atmospheric particles. When rainfall increases, the leaching of low-molecular-weight PAHs (up to four cycles) becomes more efficient (Kipopoulou 1999), this observation is in agreement with our study with the very low rainfall amounts observed during the second year. Also, the disappearance of light PAHs could result from the fact that some PAHs could have reacted with different compounds present in the air such as ozone, sulphur compounds, and nitro compounds that were captured by the rainfall and thus modify their distribution (AIRFOBEP 2004). For the rain of June 05, 2020 in Cocody, all the concentrations are lower than the detection limit of the device, except for Na which gives a value of 0.004 ng/l. This could be justified by the fact that this PAH remains the least hydrophobic compared to the others. Thus, despite its high volatility, it has a stronger affinity with the liquid phase (Chahin 2010).

The Spearman correlation coefficients are calculated for the 15 PAHs and summarized in Table 6, and the p-values in Table 5 testify to the common origin of the PAHs taken in the rainwater. The significance threshold is taken at α=0.05. Thus, the significant correlations between the sets formed by Na-PHE (r=0.72; p-value=0.02), Na-BghiP (r=0.66; p-value=0.04), BaP-IP (0.74; p-value=0.02), and BaP-Fl (r=0.76; p-value=0.01) show that these PAH could come from road traffic. Indeed, elements such as PHE, PY, BghiP, and BaP are mostly used as tracers of diesel vehicles (Khalili et al. 1995; Ravindra et al. 2006). The sets such as PHE-FLA (r=0.69; p-value=0.03), PHE-ACE (r=0.83; p-value=0.00), FLA-ACE (r=0.70; p-value=0.03), and FLA-PY (r=0.73; p-value=0.02) show a mixture of sources. Indeed, the presence of PY and PHE shows an origin linked to combustion from diesel vehicles (Ho et al. 2002; Riddle et al. 2008) and the presence of FLA and ACE which are used as markers of wood combustion and coal (Harrison et al. 1996; Zhang et al. 2008). These different correlations show the presence of multiple sources (traffic (diesel) and domestic (wood and coal)) at the sites.

In the waters collected on the two sites, the PHE/ANT ratio gives values that vary from 0 to 61.4. Two samples out of 10 collected on the Treichville site gave ratios higher than 10 (threshold), thus indicating that for these samples the presence of PAHs would be linked to the petrogenic source (Ravindra et al. 2006). The high ratio values are justified by the fact that the solubility of phenanthrene (1.29 mg/l) is higher than that of anthracene (0.073 mg/l) as well as its saturation vapour pressure, and phenanthrene has a greater facility to pass from the gaseous state to liquid and thus to dissolve in rainwater. The heavy rainfall experienced on those 2 days would have favoured atmospheric leaching of phenanthrene more than that of anthracene. The eight other samples give fairly low values and lower than 4 (threshold values for pyrolytic sources), thus justifying that the presence of these PAH would be due to the pyrolytic source. This source would then be the most important and actively contributes to the dissolution of PAHs in the collected rainwater. The BaA/BaA+CHR ratio remains lower than 0.2, which confirms that the pyrolytic source remains the source of PAHs in rainwater. Indeed, for a ratio included in the interval of 0.2–0.35, the pyrolytic source is highlighted or the average obtained for this ratio is 0.35. On the other hand, the ratio (LMW)/(HMW) remains lower than 1 in the majority on the two sites and the two seasons, confirming that the presence of PAH in water would be related to an incomplete combustion (Zhang et al. 2008).

An approach increasingly used today is based on the determination of the isotopic composition (carbon 13 (13C)), which is a conservative quantity of a molecule. The main objective of this methodology is to bring complementary information to the classical approaches of source tracking (based on the measurement of concentrations previously described). Our method consisted in comparing our 13C results with those obtained by some authors (O'Malley et al. 1996; McRae et al. 1996; Okuda et al. 2002; Peng et al. 2006) who determined 13C from PAHs from different sources in their work. The results of the isotopic analyses of two rainwater samples collected in Cocody and Treichville are recorded in Table 7.

The results of the isotopic analysis are recorded in Table 7. These results show, at the Treichville rainfall level, an enrichment in light PAHs. The 13C values for these PAHs are between −26 and −25‰. The high-molecular-weight PAHs show depletion with a ratio between −33 and −30‰. The same observation is made at Cocody where light PAHs give ratios always between −26 and −25‰, while those with high molecular weights give ratios varying from −32 to −29‰. In general, at both sites, there is depletion in the isotopic composition with an increasing molecular weight. This depletion would reflect the influence of the coal combustion source as shown by Peng et al. (2006). Indeed, these authors made the same observation by studying the sources of PAH in the cities of Zhengzhou and Urumchi. The work of Okuda et al. (2002) is also in agreement with our results insofar as these authors showed depletion in heavy PAH in the cities of Chongqui and Hangzhou by attributing that to coal combustion. In the rainfall analyzed in Cocody, the 13C of fluoranthene is estimated at −29.79 and −30.56‰ in Treichville. These values reflect depletion in this element whose main sources are road traffic and wood-burning, respectively, in Cocody and Treichville. The isotopic composition of phenanthrene is −26.08‰ in Cocody and −26.77‰ in Treichville, reflecting just like fluoranthene a source of road traffic and wood-burning. Our results are in agreement with the work of Guillon (2011) in France and Sun et al. (2003) in the United Kingdom. These authors showed that for 13C values of fluoranthene estimated at −26.5 or −27.7‰, the source would be road traffic as well as for a value of −28.1‰ for phenanthrene. These results, however, are contrary to those of Peng et al. (2006) in China who obtain varying ratios between −23 and −21‰ for fluoranthene in gasoline emissions. These differences in isotopic composition would be explained by the fact that although this composition is source related, they are also dependent on formation conditions, source material such as fuel (McRae et al. 1999) or plant (Ballentine et al. 1996).

The use of the isotope ratio allowed us to confirm the three sources of PAHs identified by the Spearman correlation matrix and the molecular indices, namely the road traffic source, the wood combustion source, and the coal combustion source.

Through this static study, we will be able to highlight the link between meteorological parameters such as air temperature (T), relative humidity (Hr), dew point (Pt C), rainfall (P), cloud cover (C), and wind speed (V) and PAHs measured in the collected rainwater. Table 8 presents the correlation coefficients obtained from the Spearman correlation matrix obtained with the XLSTAT software version 2016, with the significance threshold taken at α=0.005. Some correlations with some PAH studied during the year 2019 are obtained. Indeed, temperature and relative humidity show quite significant correlations with anthracene, phenanthrene, naphthalene, and fluoranthene which are overall low-molecular-weight PAHs. On the other hand, the condensation point shows correlations with chrysene and fluoranthene, PAHs with high molecular weight. Finally, rainfall (rain height), cloud cover, and velocity are correlated with benzo(b)fluoranthene and D(ah)A. The existing correlation between PAHs and temperature could be explained by the fact that high temperatures are conducive to aerosol formation. In fact, the solubility of PAHs in rain is strongly dependent on temperature. By passing from a temperature of 4–30 °C, the solubility of phenanthrene and anthracene is multiplied by 3.4 and 4.6 according to Schwarzenbach (1993) and Reza & Trejo (2004). In addition, variations in ambient temperature facilitate the volatilization of phenanthrene and lead to losses of certain PAHs (INERIS 2000). These losses increase with temperature (Barton et al. 1980).

Once in the air, under the effect of vaporization, the lightest PAHs can easily be captured by water droplets during washout. Albinet (2006) shows that PAHs with low molecular weight, such as PHE, ANT, and Na, tend to be in gaseous form. Gaseous and particulate PAHs are trapped in water droplets. Thus, when the rainfall increases, the share of light PAHs also increases. This can be seen in our study by the concentrations of these three elements in the rainfall of March 20, 2019 with the recorded concentrations. This observation was also made by Motelay-Massi et al. (2003).

The use of molecular indices and isotope ratios shows that the most important pyrolytic source is PAH in rainfall. A total of three pyrolytic sources were identified, namely road traffic, coal, and wood-burning.

This section highlights the health risks to which the population is exposed by being in contact with water containing PAHs. The risk assessment was done by calculating the certain index (HI and cancer risk) by applying formulas defined by the U.S. EPA and used by several authors. The result of the calculation of the ILCR, the HQ, the HI, and the cancer risk index is shown in Tables 9–12.

The estimated HQs of PAH in rainwater for dermal adsorption ranged from ND to 6.59 and from ND to 4.48 in adults and children, respectively, in 2019 (Table 8). In 2020, the quotients ranged from ND to 2.97×10⁻² for adults and from ND to 8.77×10⁻² for children (Table 8). These results indicate that HQs for both age categories in rainwater were below the U.S. EPA maximum limit of 1, excepted two samples in 2019 with some HQs more than 1. Similar results were obtained by Wei et al. (2015) and Adeniji et al. (2019). According to these authors, the probability of contracting any non-carcirogenic disease by using these water is very unlikely.

This table presents the different values obtained during the 2 years of collection on an adult and child population. The HI calculated from non-carcinogenic PAHs shows values ranging from 0 to 0.286 mg/kg/day for an adult during the first year and from 0 to 0.033 mg/kg/day during the second year. For a child, the index ranges from 0.001 to 0.844 mg/kg/day in 2019 and from 0.001 to 0.096 mg/kg/day in 2020. All of these values remain below 1, showing that the population is not at risk from these non-carcinogenic PAHs (breathing problems or skin irritation). However, the values appear to be higher for children than for adults, indicating that the former are more exposed than the adult population. A similar result was found by Olayinka et al. (2018) who show by working on the distribution of PAHs in Nigerian water samples that children would be the most susceptible to cancer due to their relatively fragile immune system compared to adults. These different calculated indexes show that although they are not consumed, these rainwaters would have a sanitary influence (skin cancer) on the population and particularly on children.

To identify the potential cancer risk to adults and children exposed, the ILCR for dermal pathway needs to be calculated based on the cancer slope factor (SF). According to some studies (Chen & Liao 2006; Petit 2016), ILCR values ≤10⁻⁶ suggest virtual safety, some values between 10⁻⁶ and 10⁻⁴ suggest potential risk, and values more than 10⁻⁴ indicate a potentially high risk. The individual ILCR assessment from PAH concentrations in water samples is presented in Table 11. Adults show an ILCR in 2019 ranged from ND to 6.03×10⁻¹, while in children, we note an ILCR ranged from ND to 1.78. In 2020, adults show some values ranged from ND to 7.16×10⁻². In children, the values are between ND and 7.97. The authors report higher values in children than in adults. Many results have been obtained from higher values of dermal ILCR. The authors report fairly high values in children than in adults. Many results have been obtained from high values of dermal ILCR. The case of Karyab et al. (2016) who reported that some dermal ILCR was almost 11 times higher than that of ingestion ILCR, and explained it by the fact that dermal exposure and its SF are so high.

In terms of cancer risk calculated with carcinogenic PAHs, the values for adults in 2019 range from 0.003 to 0.615 mg/kg/day, while in 2020, the values are lower and range from 0 to 0.113 mg/kg/day. For children, the values range from 0.008 to 1.82 mg/kg/day in 2019 and from 0 to 0.141 mg/kg/day in 2020 (Table 12).

When compared to 1×10⁻⁵ mg/kg/day (0.00001 mg/kg/day), the reference value for cancer risk is defined by the WHO. The studied waters present values of the order of 10⁻³–10⁻¹ higher than this threshold value over the 2 years, showing that particular attention should be paid to these waters used by the population. These values are almost identical to those obtained by Essumang (2010) in some surface waters of Ghana. Indeed, the author obtains values of the order of 10⁻⁴–10⁻² for the risk of skin cancer in both categories of people. This observation was made by Ekere et al. (2019), who, while studying the level of risk from PAHs in Nigeria, made the same observations on both population groups.

The PAH study revealed low values in rainwater. The total PAHs vary from 121 to 4,204 ng/l in 2019 with a strong contribution of phenanthrene in the dry period and BaP in the rainy period. In 2020, the total varies from 0.4 to 318.3 ng/l with a strong contribution of PI in the dry period and BaP in the rainy period. The study of the various sources shows that the pyrolytic sources (road traffic, combustion of wood and coal) remain the most important with a predominance of light PAHs in Treichville and heavy PAHs in Cocody. The health risk assessment gives indices of the order of 10⁻³–10⁻¹ in 2019 and 10⁻⁴–10⁻² in 2020, showing that the populations would be exposed to a risk of skin cancer.

This study was possible in close collaboration with the Laboratory of Atmospheric Physics and Fluid Mechanics (LAPA-MF), which provided the rain collectors and air filters acquired under the multisectoral assistance programme (PASMU). The analyses were financed by the Strategic Support Programme for Scientific Research (PASRES). We thank the Stable Facility laboratory of the UC DAVIS University of California, which kindly carried out the isotopic analyses.

All relevant data are available from an online repository or repositories.