Gaseous Pollutants and Particulate Matter in Ambient Air: First Field Experiment in an Urban Mediterranean Area (Nador, Morocco)

Received: December 13, 2022
Revised: May 21, 2023
Accepted: May 23, 2023

 Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Cite this article:

Oujidi, B., Benchrif, A., Tahri, M., Zahry, F., Bounakhla, M., Bazairi, H., Mhammdi, N., Snoussi, M. (2023). Gaseous Pollutants and Particulate Matter in Ambient Air: First Field Experiment in an Urban Mediterranean Area (Nador, Morocco). Aerosol Air Qual. Res. 23, 220451. https://doi.org/10.4209/aaqr.220451

HIGHLIGHTS

• First investigation of urban air quality in Nador city.
• NO₂, O₃, and CO levels were lower than the national standard limits.
• Dominance of anthropogenic emissions shown.
• Nador city had higher biomass black carbon levels compared with other areas.

Keywords: Western Mediterranean Basin, Measurement campaigns, PM10, PM2.5, Black carbon

1 INTRODUCTION

Over the past decade, several studies have reported that exposure to air pollutants, such as gaseous and particulate matter, can increase mortality and morbidity by causing respiratory and cardiovascular disease in affected populations (Nguyen et al., 2022; HEI, 2019; Pattenden et al., 2010; Pope III and Dockery, 2006). Ambient air pollution can also have a negative influence on the climate (Bodor et al., 2020; Cholakian et al., 2019; Jeong et al., 2011), and it has been identified as causing a variety of health-related effects, including cancer (Bennett et al., 2019; Boffetta and Nyberg, 2003). For example, PM₁₀ and PM_2.5 can easily become trapped in the alveoli of the lungs, leading to lung damage and potentially premature death (Das et al., 2015; Pérez et al., 2008; Dockery and Pope, 1994; Dockery et al., 1993). Black carbon is a component of fine particulate matter that can act as a universal transporter for a large variety of toxic chemicals, taking them to the lungs, the body’s major defense cells, and possibly the systemic blood circulation (Janssen et al., 2012).

Rapid urbanization and socioeconomic development without accompanying environmental protections have exacerbated air pollution (Yang et al., 2020). The main sources of this pollution include industrial emissions, power generation, transportation, residential combustion, and the open burning of crop residues and other solid waste (Das et al., 2022; Turner et al., 2020; Das et al., 2020, 2018; Nagpure et al., 2015). It is therefore crucial to quantify how the gaseous pollutants, particulate matter and its chemical composition, and the black carbon may impact the climate and human health.

In Morocco, the first assessments of air quality in cities started in the 2000s, with studies carried out in Casablanca and Mohammedia showing a significant relationship between levels of air pollution and damage to public health (Ministry of Territorial, Urban, Housing and Environment Planning and Ministry of Health, 2000, 2003). Moreover, Al Wachami et al. (2022) reported that recent studies have argued the effects of air pollution on the development and intensification of respiratory diseases in Morocco. For instance, Abrouki et al. (2021) identified a statistically positive correlation between SO₂ and NO₂ pollutants and infant asthma for Casablanca, Mohammedia, Agadir, Marrakech, Salé, and Fez, while Gryech et al. (2022) reported that the high prevalence of respiratory diseases in Rabat was related to air pollution. In 2018, the government set up and implemented a national air program (PNAIR) to (1) reinforce the national air quality monitoring network especially in big industrial cities and those with traffic pollution concerns, such as Casablanca, Mohammedia, and Rabat, (2) reduce the emission of pollutants generated by the transport and industry sectors, strengthen the regulatory framework for air pollution, and (3) improve communication and awareness (Ministry of Energy, Mines and Sustainable Development, 2018). In addition, further research has assessed the air quality in several other Moroccan cities (Tahri et al., 2022a, 2022b; Bounakhla et al., 2022; Otmani et al., 2020; Benchrif et al., 2018; Tahri et al., 2017, 2013, 2012; Bounakhla et al., 2009). Concerning the coastal Mediterranean cities of Morocco, only the western portion has been the subject of a recent study of air pollution (Benchrif et al., 2018, 2022), despite the Mediterranean basin being sensitive to both climate change and air pollution, particularly in the summer (Doche et al., 2014).

The Nador city is located on the eastern portion of Morocco’s Mediterranean coast, and this first study of this area’s air pollution aims to characterize the concentration of the gaseous pollutants NO₂, CO, and O₃. It also investigates the effect of meteorological variables on temporal variations in air pollutant concentrations using a multi-univariate regression approach and attempts to characterize the two main forms of particulate matter (PM₁₀, PM_2.5) and their elemental compositions. This study also assesses the black carbon in fine particles and detects anthropogenic emissions through the enrichment factors and the elemental ratios.

 
2 DATA AND METHODS

 
2.1 Study Area

The city of Nador and its watershed are located on the shores of the Marchica lagoon (Fig. 1), and it has undergone an urban boom in recent decades (Aitali et al., 2022), with it having almost 400,000 inhabitants according to the 2014 census (HCP, 2015). Nador has a Mediterranean climate with wet and dry seasons, with hot, dry summers contrasting with mild, rainy winters. Socioeconomic activities around the city comprise about 150 industries located in an industrial area within 20 km. These mostly comprise metallurgical activities with some agro-industrial, chemical, and para-chemical units. Most were built before 2003, when the law for impact studies in Morocco forced industries to set up a system for pretreating solid, liquid, and gaseous pollutants before they are discharged into the external environment (Oujidi et al., 2021a). Other polluting activities include (1) agriculture on the Bouarg plain that excessively applies fertilizers and pesticides and tourism activity around the Marchica lagoon, about 500 m from our study area, as well as (2) atmospheric emissions from industry and waste incineration, which took place for several decades before a controlled landfill was established in 2016 (Oujidi et al., 2020).

Fig. 1. (A) Location of the sampling site in an urban area in Nador city, (B) located in the Mediterranean region, and (C) distance from the Marchica lagoon and in an urban area.

 
2.2 Methods

2.2.1 Instrumentation and sampling

All the equipment used to analyze gaseous pollutants, particulate matter, black carbon, and elemental composition belonged to the National Centre for Nuclear Energy, Science and Technology (CNESTEN). Gaseous pollutants were measured through the first two daily sampling campaigns of four days each from October 10–13, 2016 and from April 5–8, 2017. Nitrogen dioxide (NO₂) was measured every half hour using a Chemiluminescence technique (Horiba, APNA-370) with a flow rate of 0.8 L min^–1 and a minimum sensitivity detection of 0.5 ppb for concentrations lower than 0.2 ppm. Ozone (O₃) was also measured every half hour using a UV photometry technique with an automatic analyzer of type O341M (Environment SA) with a flow rate of 1.5 L min^–1 and a minimum sensitivity detection of 0.4 ppb. Carbon monoxide (CO) was analyzed every quarter of an hour using non-dispersive infrared analysis by an automatic analyzer of type APMA-370 (HORIBA) with a flow rate of 1.5 L min^–1 and a minimum sensitivity detection of 0.05 ppb for concentrations lower than 10.0 ppm. The performance features of the gaseous pollutant measurements were evaluated according to the following standards: NM EN 14211:2012 for NO, NO₂, NO_x (IMANOR, 2012a), NM EN 14626:2012 for CO (IMANOR, 2012b), and NM EN 14625:2012 for O₃ (IMANOR, 2012c).

Total PM₁₀ (aerosols < 10 µm in diameter) and fine PM_2.5 (aerosols < 2.5 µm in diameter) measurements were taken over a two-week period during the dry season from July 17 to July 30, 2018. The PM cutoff cyclone manufactured by URG^® was operated at 1 m³ h^–1 (approx. 16.7 L min^–1) and used to collect fine and total particles on 47-mm diameter PTFE (Whatman®) filters (2 µm pore size) for a sampling period of 24 hours. The loaded filters were placed in a desiccator for 48 hours before being weighed to minimize the effect of particles capturing moisture from the atmosphere (Benchrif et al., 2022). PM₁₀ and PM_2.5 concentrations were calculated as the ratio of the measured net weight for each filter compared to the volume of air sampled.

All the PM samplers and gaseous pollutant analyzers were located on the terrace roof of the Provincial Directorate of Equipment, Transport, Logistics and Water (DPETLE) (Fig. 1) (35°10′27.57′′; 2°55′32.56′′). This was 8 meters above ground level in the city center of Nador, thus ensuring that there were minimal obstacles to airflow that could affect PM concentrations.

 
2.2.2 Chemical analysis

The collected PM samples were digested with 8 ml of concentrated HNO₃ and 2 mL of H₂O₂ in a household microwave oven (Milestone MAXI14) at 200°C for 45 minutes. The digested samples were then kept in a refrigerator at 4°C for later analysis. The extracted solutions were analyzed using two techniques, namely total reflection X-ray fluorescence (TXRF) and microwave plasma atomic emission spectrometry (MP-AES).

For the TXRF analysis, gallium was added as an internal standard for quantifying the elemental concentrations. Some 10 µL of the digested filter sample was pipetted onto a siliconized quartz glass carrier and dried under an infrared lamp. The sample was then analyzed by a TXRF type S2-Picofox from Bruker to measure P, K, Ca, V, Cr, Cu, Fe, K, Mn, Ni, Zn, Sr, Pb, and Ba concentrations.

The Al and Na concentrations were measured using MP-AES, with a blank filter being analyzed using the same procedure to monitor background contamination. The blank values were then subtracted from the analyzed values. The accuracy of this analysis was evaluated using a standard reference material (SRM, 2783 Air particulate on Filter Media- NIST), and the relative deviation of the obtained concentrations from the reference values varied from 2% to 14% in absolute value.

 
2.2.3 Black carbon measurement

Equivalent black carbon (BC) and light-absorbing carbon (LAC) in PM_2.5 samples were analyzed for the dry period using a multi-wavelength absorption black carbon instrument (MABI) (Manohar et al., 2021). This measured LAC at seven different wavelengths (405 nm, 465 nm, 525 nm, 639 nm, 870 nm, 940 nm, and 1050 nm). To differentiate contributions from biomass burning (BBC) from those of burning fossil fuels (FBC), three wavelengths were considered: 639 nm for BBC and the difference between the concentrations at 405 nm and 1050 nm for FBC (Mutahi et al., 2021; Tuso et al., 2020).

 
2.3 Data Processing

2.3.1 Enrichment factor

The enrichment factor (EF) is often used to detect the anthropogenic influence by distinguishing it from the natural background metals in atmospheric elemental levels. The EF is determined by normalizing an observed element to the crustal concentration of a reference element. Generally, aluminum (Hsu et al., 2016) and iron (Enamorado-Báez et al., 2015; Lee and Hieu, 2011) are often used for normalization to the crust because these two elements are stable in the soil and essentially originate from natural sources (Barbieri, 2016; Ackermann, 1980). In this study, iron was used for normalization, with the reference values (Upper Continental Crust) being taken from Rudnick and Gao (2003). The EF values were calculated according to the work of Zoller et al. (1974), as shown in Eq. (1):

where (C_n/C_Fe)_Sample and (C_n/C_Fe)_Crust represent the concentration ratios between the target metal (n) and the reference metal (Fe) in the PM_2.5 or PM₁₀ samples and the crust, respectively. By convention, an arbitrary average EF value less than 10 can be regarded as indicating that a trace metal in an aerosol has a significant crustal source, with these being termed non-enriched elements. An EF value between 10 and 100, meanwhile, indicates that the element has been moderately enriched, while an EF value greater than 100 indicates that a significant proportion of the element has come from a non-crustal source, with these being referred to as anomalously enriched elements (Tahri et al., 2017; Das et al., 2015).

2.3.2 Meteorological data

Meteorological data for air temperature, precipitation, relative humidity, wind speed, wind direction, and pressure were obtained during the sampling periods from weather station Aroui (35-09N, 002-55W), which was located about 26 km from the study site. The R package “worldmet” (Carslaw and Ropkins, 2012) was used to retrieve meteorological observations from the NOAA Integrated Surface Database (ISD) (available online at: https://www7.ncdc.noaa.gov/CDO/cdo).

 
2.3.3 Multivariate linear regression analysis

A multivariate linear regression (MLR) analysis was performed to assess the variation of each examined pollutant on taking into account the meteorological impact. This approach is frequently applied to model the correlation between pollutant concentration (as a response variable) and meteorological parameters (as independent variables). MLR uses the least-squares algorithm to estimate the regression coefficients (r²) between actual and predicted values. The extent of each pollutant’s presence is therefore modeled by the linear regression equation (Bounakhla et al., 2023; Luo et al., 2021) presented in Eq. (2):

where Pollutant is the response variable, b₁ represents the regression coefficients for the impact of the corresponding meteorological variables on the pollutant level, x_i are the independent variables, and C is the intercept of the model.

 
2.3.4 Statistical analysis

Descriptive statistics (min, max, mean, and standard deviation) were calculated for the gaseous pollutants, the particulate matter, the chemical composition, the black carbon, and the enrichment factor using the Excel software. The statistical software programming language R (R Core Team, 2020) with its “Openair” package (Carslaw and Ropkins, 2012) were used to plot time variations for the gaseous pollutants. The Spearman’s rank coefficient was calculated using IBM’s SPSS statistics software version 20 to assess the relationship between the particulate matter and climatic conditions in the air of Nador.

 
3 RESULTS AND DISCUSSION

 
3.1 Levels of Gaseous Pollutants, BC, PM₁₀, and PM_2.5 together with Meteorological Data

A summary of the average concentrations of PM₁₀, PM_2.5, BC, NO₂, CO, and O₃, as well as the meteorological data at the measurement site in Nador are given in Table 1. This table includes information about the research period, measuring techniques and methods, and the reported concentration values. The meteorological conditions of Nador were characterized by low autumn (October 2016) temperatures varying from 7.8 to 26°C with a mean value of 15.7°C and spring (April 2017) temperatures ranging from 7.1 to 23°C with a mean value of 15°C, compared with high temperatures in summer (July 2018) that varied from 17 to 36°C with a mean value of 26°C. The sampling periods were also characterized by the absence of rainfall. The relative humidity varied between 29.8% and 100% in autumn with a mean of 68.3% and between 43.8% and 100% in spring with a mean of 75.4%, compared with summer values of 14% to 83.2% with a mean of 44.7%. The wind speed varied between 0.5 and 7.7 m s^–1 in autumn with a mean of 2.4 m s^–1, between 0.5 and 9.8 m s^–1 in spring with a mean of 3.9 m s^–1, and between 0.5 and 8.8 m s^–1 in summer with a mean of 2.6 m s^–1.

 
3.2 The Gaseous Pollutants’ Concentration Patterns

3.2.1 Variation in the gaseous pollutants

The results of the descriptive analysis of gaseous pollutants in Nador are summarized in Table 1. Fig. 2 presents the hourly variation in O₃, NO₂ and CO concentrations, while the daily variations are illustrated in Fig. S1. The NO₂ concentrations vary between 8.1 and 66.8 µg m^–3 with an average of 29.7 µg m^–3 during autumn and between 0 and 52.6 µg m^–3 with an average of 20.4 µg m^–3 during spring. The O₃ concentrations vary between 2.1 and 32.1 µg m^–3 with an average of 14.7 µg m^–3 during autumn and between 8.6 and 66.3 µg m^–3 with an average of 44.2 µg m^–3 during spring. Finally, the CO concentrations varied between 0.01 and 1.2 mg m^–3 with an average of 0.3 mg m^–3 during autumn and between 0.1 and 0.5 mg m^–3 with an average of 0.3 mg m^–3 during spring.

Fig. 2. Hourly variation of O₃, NO₂ and CO concentrations in Nador city measured during four days sampling experience (from 10 to 13 October 2016 and from 05 to 08 April 2017). Boxplot indicates minimum, maximum, median (dash line), and 25/75% quartiles (boxes) in the dataset.

The 98^th percentile of the NO₂ hourly average and the eight-hourly average of O₃ during the two measurement campaigns were 51.1 µg m^–3 and 31.3 µg m^–3, respectively, which are below the national standard values of 200 µg m^–3 and 110 µg m^–3, respectively. As shown in Fig. 2, the maximum NO₂ peaks recorded during the two measurement campaigns occurred in daytime at the beginning of the day and the end of the afternoon. However, the opposite trend was noted for O₃, mostly due to photochemical formation, weather conditions, and road traffic activities associated with NO₂ emissions (Brancher, 2021; Hashim et al., 2021; Wu and Xie, 2017). For example, an increase in global solar radiation and the height of the boundary layer between early morning and midday causes a drop in NO₂ concentration and an increase in O₃ concentration (Yu et al., 2021; Liu et al., 2022). There was also a clear difference in NO₂ and O₃ concentrations between weekdays and weekends, with the NO₂ level being higher on weekdays than during the weekend due to traffic activities, while the level of O₃ was higher at weekends than on weekdays. Han et al. (2011) reported that this pattern of temporal variability can also be found in other locations, with them positing various hypotheses to explain the observed weekend effects. Firstly, ozone formation is sensitive to concentrations of volatile organic compounds (VOCs) and NO₂. Berezina et al. (2020) reported that various studies have documented different ozone responses to changes in precursor emissions, with them depending mostly on the local VOC-to-NO_x ratio in the atmosphere. For example, at very low levels of VOC/NO_x (<< 10 ppb C (part per billion carbon) to ppb NO_x), ozone production will increase alongside a decrease in NO_x emissions. At high conditions (> 10), NO_x becomes the limiting reactant and ozone formation becomes more efficient with the addition of NO_x (Wolff et al., 2013). With the observed reduction in NO_x emissions over the weekend being about three times greater than the reduction in VOCs, the weekends have higher VOC/NO_x ratios than the weekdays, leading to enhanced ozone formation. Secondly, the NO_x concentrations are lower at the weekend due to lower levels of traffic and the related emissions, so the inhibitory effect of NO_x on O₃ formation is weaker and more O₃ is produced.

The daily maximum for the 8-hour moving average of CO during the two measurement campaigns was calculated as 0.47 mg m^–3, well below the national standard value of 10 mg m^–3 (Fig. 2). The CO level is a reflection of road traffic pollution (Das et al., 2022; Zhao et al., 2019; Nagpure and Gurjar, 2012; Wu and Wang, 2005), so it is unsurprising that the diurnal cycle of CO concentration shows small double peaks in the morning and evening. However, the amplitude ratios for daytime compared to nighttime were found to be close to unity (ranging from 0.98 to 1.26), suggesting stable CO levels. In addition, the Fig. 2 shows that the concentrations of NO₂ and CO were strongly negatively correlated with O₃. Higher concentrations of NO₂ and CO are associated with lower O₃ concentrations and vice versa due to photochemical reactions (Kovac-Andric et al., 2013).

 
3.2.2 Multivariate linear regression analysis

Our model analysis used the collected hourly data for the gaseous pollutants NO₂, CO, and O₃. In total, 200 observations had been collected and distributed over the measurement period. Our comprehensive evaluation of this collected data was based on the processing steps summarized by Bounakhla et al. (2022). Once selected, the model needed to be interpreted by analyzing all the statistical parameters, including the root mean square error (RMSE) and R-squared (R²). The best-fitting models are characterized by a low RMSE, a high R-squared, and a small difference between the RMSE of the training set and the test set. The obtained test, training, and cross-validation (CV) root mean square error and correlation coefficient (r²) values are presented in Table S1. Several multiple regression models were examined to determine the best-fitting model, and it was found that fitting the non-linear model to the meteorological variables yielded the best results. The proposed regression equation for the relationship between the pollutants (NO₂, CO, and O₃) and the meteorological variables is expressed in Eqs. (3–5). Two meteorological parameters (wind speed and temperature) have a significant impact on NO₂ and CO concentrations. Wind speed has a negative effect on NO₂ and CO concentrations, while temperature has a positive effect. This is because lower wind speeds make it more likely that pollutants will accumulate, while a higher temperature causes more pollutants to become trapped under a shell of warm air and thereby accumulate. He et al. (2017) observed similar results. For the O₃ concentration, all the considered meteorological parameters (wind speed, temperature, wind direction, and relative humidity) have a significant influence. For instance, a decrease in wind speed, temperature, or relative humidity leads to a decline in O₃ concentration. Similar findings were also reported by Liu et al. (2020) and Yang et al. (2021).

Fig. 3. (a) Biomass burning contribution (BBC) and (b) Fossil fuel contribution (FBC) in Nador city.

 
3.3 Black Carbon in Fine Particles

The contributions of burning biomass (BBC) and fossil fuels (FBC), established using the MABI optical measurement, are shown in Fig. 3. The concentrations of BBC and the FBC ranged from 3.1 to 8.4 µg m^–3 and 0.9 to 1.8 µg m^–3, respectively, so biomass burning made a much larger contribution to black carbon concentrations than the burning of fossil fuels. This can be explained by how wood burning is usually used in traditional hammams and artisan bakeries in Nador (Otmani et al., 2022; Ryś and Samek, 2021; Ielpo et al., 2020). Lourenço et al. (2018) also reported that frequent fires occur in the Gourougou forest, which is located about 10 km from the sampling site, which are likely to contribute to biomass black carbon levels.

Fig. 4. Variation of particle matter PM₁₀ and PM_2.5 with climatic parameters (Temperature, wind speed (WS), and relative humidity (RH)) in Nador city during dry sampling periods.

 
3.4 The Variation in PM₁₀ and PM_2.5 Concentrations

The results of the descriptive analysis of particulate matter in Nador are given in Table 1. The concentration of PM₁₀ during the dry season ranged from 24.4 to 57.3 µg m^–3 with a mean value of 41.4 µg m^–3 (Fig. 4). The 90.4^th percentile of daily means for PM₁₀ was 52.6 µg m^–3, with three daily values exceeding the national daily health-exposure limit of 50 µg m^–3. The concentration of PM_2.5 during the dry season ranged from 13.5 to 32.9 µg m^–3 with a mean value of 22.2 µg m^–3 (Fig. 4). The 99^th percentile of daily means for PM_2.5 was 32.4 µg m^–3, with almost all the daily concentrations exceeding the international standard limit of 15 µg m^–3 (WHO, 2021). The ratio of PM_2.5/PM₁₀ varied between 0.3 and 0.8 with a mean value of 0.5. A higher PM_2.5/PM₁₀ ratio (between 0.5 and 0.8) indicates there is a major contribution from fine anthropogenic particles. Climatic conditions generally affect the transport and distribution of particulate matter, so the correlation between particulate matter and climatic conditions (temperature, wind speed, and relative humidity) was examined using Spearman’s correlation analysis. Nevertheless, no significant correlation was found between particulate matter and climatic parameters.

 
3.5 Elemental Composition

The elemental compositions of the particulate matter are given in Table 2. Sixteen elements were identified in the PM_2.5 and PM₁₀ samples, with these being Na, K, Ca, P, Al, Fe, Mn, Ti, Ba, V, Sr, Ni, Cu, Zn, Cr, and Pb.

For PM₁₀, the average elemental concentrations (ng m^–3) during the sampling periods were as follows: Ca (8150) > Na (1380) > Al (1190) > Fe (550) > K (250) > P (190) > Zn (100) >Ba (46) > Ti (24) > Cr (18) > Cu (11) > Mn (10) > V (9) > Pb (8) > Ni (7) > Sr (4).

For PM_2.5, the average elemental concentrations (ng m^–3) during the sampling periods were as follows: Ca (6590) > Na (990) > Al (860) > Fe (430) > K (270) > P (210) > Zn (120) > Cu (47) > Ba (27) > Cr (19) > Ti (15) > Pb (11) > V (10) > Ni (9) > Mn (6) > Sr (3).

A comparison of the elemental compositions of PM_2.5 and PM₁₀ (Fig. S2) reveals that the concentrations of elements from anthropogenic sources (Pb, Cu, Zn, Cr, Ni and V) were higher in PM_2.5, while the major elements related to soil dust (Ca, Na, Al, Ti, Mn, Fe, Sr, and Ba) were higher in PM₁₀. Similar findings have been reported by Tahri et al. (2017) for Kenitra, Morocco.

 
3.6 Comparison with Other Studies

A comparison of the obtained mass concentrations in Nador with other locations in Morocco and the Mediterranean Basin are summarized in Table 3. This shows that the average concentrations of fine particle matter and their trace elements are mostly lower in Nador than they are in the cities of Kenitra (Atlantic coast) and Meknes (inland), although Cu and Zn concentrations were higher in Nador than Meknes (Tahri et al., 2013; Bouh et al., 2013). The comparisons with Barcelona, Marseille, and Genoa reveal that the average PM_2.5 concentrations and their trace elements are higher in Nador city (Salameh et al., 2015). Cu, Zn, Cr, and Ni concentrations are higher than those of Venice, and Zn concentration was higher than in Thessaloniki (Salameh et al., 2015), while Cu concentration was found to be higher than in Cairo (Shaltout et al., 2020). The average BC concentrations were higher than those recorded in other Morocco cities, such as Salé (Otmani et al., 2022), Kenitra (Bounakhla et al., 2022), and Tetouan (Benchrif et al., 2018, 2022), as well as other Mediterranean cities like Cairo and Athens (Shaltout et al., 2020; Kalogridis et al., 2018).

 
3.7 Characterization of Anthropogenic Emissions

Two approaches were used to characterize anthropogenic emissions according to enrichment factors and elemental ratios.

3.7.1 Enrichment factors

For both PM_2.5 and PM₁₀, the average enrichment factors (EFs) were less than 10 for Ti, Mn, Sr, K, Al, Na, Ba, and V, indicating that these trace metals in the aerosols have a significant crustal source, so they are non-enriched elements. The EFs for Ni, Cr, P, Ca, Cu, and Pb were between 10 and 100, indicating that these elements were moderately enriched. However, only Zn showed an EF higher than 100, indicating that a significant proportion of this element had come from a non-crustal source, so it is an enriched element in Nador city (Fig. 5) (Hamdan et al., 2021; Wu et al., 2019; Feng et al., 2016; Tahri et al., 2017; Das et al., 2015). The high levels of enrichment for Zn, Pb, and Cu indicate that these trace elements are the main toxic anthropogenic elements in the PM of Nador. Similar results were found for sediment samples from the Marchica lagoon and its watershed around Nador (Oujidi et al., 2021a, 2021b). It is therefore crucial to identify the potential sources that are contributing to the high concentrations of these elements.

Fig. 5. Enrichment factors of the measured elements in PM₁₀ and PM_2.5.

3.7.2 Elemental ratio

Trace elements can be emitted into the atmosphere from a variety of natural sources, such as volcanoes, dust storms, forest fires, and sea spray, as well as various anthropogenic sources, such as fossil fuel and biomass combustion, iron and steel production, road traffic, shipping emissions, cement production, and waste incineration (Habeebullah et al., 2022; Das et al., 2015; Tahri et al., 2017). To estimate the potential sources in Nador city, elemental ratio methods were used as key diagnostic tools (Benetello et al., 2018; Arditsoglou and Samara, 2005; Ayrault et al., 2010). The anthropogenic elemental ratios for (1) Fe/Al; (2) Cu/Pb and Cu/Zn; (3) Zn/Pb; and (4) V/Ni in Nador city were compared with those of previous studies, as summarized in Table 4. It was found that:

• The Fe/Al ratios varied between 0.46 and 0.50, respectively, for PM₁₀ and PM₅ in Nador, which is similar to the value from crustal soil (0.4) (Chifflet et al., 2018; Reimann and de Caritat, 1998). Fe/Al mass ratios are usually used to identify natural inputs from the resuspension of terrigenous aerosol, but a high Fe/Al ratio suggests that Fe has been enriched by anthropogenic combustion sources (Mehra et al., 2020).
• The Cu/Zn and Cu/Pb ratios were found to be 0.1 and 1.4, respectively, in PM₁₀ and 0.4 and 4.7 in PM₅, respectively. With the exception of the Cu/Pb ratio for PM_2.5, the results are similar to those found in the urban Veneto region of Italy (Cu/Zn ratios of 0.1–0.4), which were mainly attributed to traffic emissions (Benetello et al., 2018). For the Cu/Pb ratio in PM_2.5, a similar ratio (4.6) was found in Turin, Italy, with this also being attributed to traffic pollution (Ziegler et al., 2021), while a Cu/Pb ratio of 2.38 was found at traffic locations in England, Wales, and Scotland in the United Kingdom (Font et al., 2015).
• The Zn/Pb ratios were 12.6 and 12 in PM₁₀ and PM₅, respectively, which is similar to the Zn/Pb ratio found for construction dust (11.3) in Fushun, China (Kong et al., 2011), indicating that this source of pollution could explain our Zn/Pb ratios. Previous studies have indicated that high Zn/Pb ratios were found for diesel vehicle emissions (7.6), scrap metal incineration (8.4), and cement plants (42) in Thessaloniki, Greece (Arditsoglou and Samara, 2005; Samara et al., 2003). Very high Zn/Pb ratios were also found for construction dust (1537.2) in the city of Raipur, India (Matawle et al., 2015).
• The V/Ni ratios were 1.2 and 1.1 in PM₁₀ and PM₅, respectively. A similar ratio was found in the ambient PM_2.5 of the coastal city of Gdynia, Poland (1.19), potentially due to local mixed industrial sources with possible contributions from shipping activities in the port area (Siudek, 2020).

 
4 CONCLUSIONS

We carried out the first assessment of the urban air quality in the city of Nador through three short-term campaigns to monitor gaseous pollutants (NO₂, CO, and O₃) and particulate matter (PM₁₀ and PM_2.5) between 2016 and 2018. Our results indicate that all gaseous pollutant concentrations were below the national standard limits. Nevertheless, Zn and Cu were highly enriched in PM_2.5, while in PM₁₀, Zn was highly enriched and Cu and Pb were moderately enriched, indicating the presence of anthropogenic emission sources in the city of Nador. Furthermore, the high PM_2.5/PM₁₀ ratio also indicates a significant contribution from anthropogenic sources to PM, while the elemental ratio analysis points to road traffic as being the largest anthropogenic source. The distribution of BC indicates that the contribution from burning biomass, such as in traditional hammams and artisan bakeries, was more significant than that from burning fossil fuels. Moreover, the multivariate analysis revealed that the relationships between NO₂, CO, O₃ concentrations and meteorological variables are not linear. Although wind speed has a negative effect on both NO₂ and CO concentrations, temperature has a positive effect. Overall, the establishment of an air-quality management plan for Nador is recommended, and this should include a long-term study and the establishment of a network of monitoring stations in order to identify and control activities that degrade air quality. Nevertheless, this study is limited by a small sample size and the use of a single sampling station, so future studies should consider using more observation sites and recording data over longer periods to analyze trends in airborne pollutants’ concentrations.

 
ACKNOWLEDGMENTS

The author thanks all partners of the Lag-Nad 2016 project that supported this study. Thanks to Dr Mohammed El Bouch, Director of National Laboratory for Studies and Pollution Monitoring (LNESP) Morocco; Dr Mostafa Layachi, Researcher at National Institute of Fisheries Research (INRH) Morocco; and Dr Rachid Bouchnan, Professor at Abdelmalek Essaâdi University ENS-Tétouan, Morocco for their precious help. Many thanks to the Provincial Directorate of Equipment, Transport, Logistics and Water (DPETLE) of Nador for the use of terrace. We are grateful to the editors and reviewers who contributed to the improvement of the quality of this manuscript.