Aerosols in Northern Morocco (Part 4): Seasonal Chemical Signatures of PM_2.5 and PM₁₀

Abstract

Atmospheric aerosols are recognized as a major air pollutant with significant impacts on human health, air quality, and climate. Yet, the chemical composition and seasonal variability of aerosols remain underexplored in several Western Mediterranean regions. This study presents a year-long investigation of PM_2.5 and PM₁₀ in Tetouan, Northern Morocco, where both local emissions and regional transport influence air quality. PM_2.5 and PM₁₀ samples were collected and analysed for total mass and comprehensive chemical characterization, including organic carbon (OC), elemental carbon (EC), water-soluble ions (WSIs), and sugar tracers (levoglucosan, arabitol, and glucose). Concentration-weighted trajectory (CWT) modelling and air mass back-trajectory analyses were used to assess potential source regions and transport pathways. PM_2.5 concentrations ranged from 4.2 to 41.8 µg m⁻³ (annual mean: 18.0 ± 6.4 µg m⁻³), while PM₁₀ ranged from 11.9 to 66.3 µg m⁻³ (annual mean: 30.8 ± 9.7 µg m⁻³), with peaks in winter and minima in spring. The PM_2.5-to-PM₁₀ ratio averaged 0.59, indicating a substantial accumulation of particle mass within the fine fraction, especially during the cold season. Carbonaceous aerosols dominated the fine fraction, with total carbonaceous aerosol (TCA) contributing ~52% to PM_2.5 and ~34% to PM₁₀. Secondary organic carbon (SOC) accounted for up to 90% of OC in PM_2.5, reaching 7.3 ± 3.4 µg m⁻³ in winter. WSIs comprised ~39% of PM_2.5 mass, with sulfate, nitrate, and ammonium as major components, peaking in summer. Sugar tracers exhibited coarse-mode dominance, reflecting biomass burning and biogenic activity. Concentration-weighted trajectory and back-trajectory analyses identified the Mediterranean Basin and Iberian Peninsula as dominant source regions, in addition to local urban emissions. Overall, this study attempts to fill a critical knowledge gap in Southwestern Mediterranean aerosol research by providing a comprehensive characterization of PM_2.5 and PM₁₀ chemical composition and their seasonal dynamics in Tetouan. It further offers new insights into how a combination of local emissions and regional transport shapes the aerosol composition in this North African urban environment.

1. Introduction

The Mediterranean basin has been well documented as one of the most sensitive regions in the world in terms of climate and air quality [1]. Its location at the cross-roads of air masses transporting both natural and anthropogenic aerosols from Europe, Asia, and Africa results in a complex interplay of atmospheric influences [2]. The diverse range of aerosol sources, combined with the region’s intricate orography and atmospheric dynamics, gives rise to a complex mixture of components, including organic matter, inorganic salts, elemental carbon, trace metals, etc., [3]. The specific chemical profiles of these aerosols vary depending on their source regions and transport pathways to inland areas [4]. Aerosols, which have a relatively short atmospheric lifetime and originate primarily from nearby regions, also play a crucial role in shaping the Mediterranean climate. As a result, they constitute a significant regional climate forcing of local origin [5].

Understanding the chemical composition of size-segregated aerosols is important for assessing their impacts on climate, air quality, and human health [6,7]. Aerosols, commonly referred to as suspended particulate matter (PM) in the atmosphere, play a pivotal role in these domains [8,9]. PM is typically classified based on its aerodynamic diameter, with PM₁₀ and PM_2.5, representing particles with an aerodynamic diameter less than or equal to 10 and 2.5 µm, respectively, being the most commonly monitored particle sizes [10]. This classification is crucial, as larger particles tend to settle more quickly and impact air quality on a local scale, whereas smaller particles remain suspended longer, traveling across regions and even continents. Both PM_2.5 and PM₁₀ fractions have been extensively studied over the Mediterranean region using ground-based measurements and satellite retrievals, with a focus on both the Western and Eastern Mediterranean areas [11,12,13,14]. Fine aerosols are largely dominated by particulate organic matter and secondary inorganic aerosols, notably sulphate, as expected based on their sources and formation mechanisms [3,15,16]. In contrast, the coarse fraction is primarily composed of mineral dust, along with some water-soluble ions such as particulate nitrate [17,18]. Overall, a common finding across studies is that aerosols in the Mediterranean basin—where they originate from a complex mix of natural and anthropogenic sources—vary significantly in their chemical composition depending on particle size, source region, and atmospheric transport processes [11].

Studies across the Mediterranean highlight the complex interplay between local and transported pollution [19,20]. They reveal transboundary aerosols frequently dominate urban PM levels, complicating local air quality management [19]. Cyprus exemplifies this pattern, where regional sources contribute 53–61% of fine PM annually, though winter shows a stronger local influence (≤70% PM_2.5 from domestic biomass burning) [19]. This contribution is strongly seasonal: winter PM is largely dominated by local emissions, particularly biomass burning for domestic heating under stagnant meteorological conditions, while summer is influenced by long-range transport and local dust resuspension. Similar patterns have been observed across the Eastern and Central Mediterranean, where air masses arriving in Athens predominantly originate from northern sectors (Central/Eastern Europe and Western Turkey), reaching ~85% during summer [21]. In the Western Mediterranean, studies in Tetouan (Northern Morocco) demonstrated that long-range contributions mainly arise from the Mediterranean Basin, including the densely populated coasts of Spain and North Africa, with a notable influence of maritime traffic on sulfate-rich aerosols, while local sources are dominated by urban emissions [18].

This study presents one year of aerosol (PM_2.5 and PM₁₀) mass measurements conducted in an urban Mediterranean environment in Tetouan, Northern Morocco. To the best of our knowledge, these represent one of the few long-term aerosol datasets in the Southwestern Mediterranean, a region previously identified as understudied with respect to dominant aerosol types and mixing states [12,14,15,16,22]. Combining PM_2.5 and PM₁₀ mass measurements with chemical analysis enables better discrimination between natural and anthropogenic sources and quantification of their respective contributions. This, in turn, provides essential baseline data for developing targeted air quality management and pollution control strategies. Previous work in Northern Morocco under the EGIDE/VOLUBILIS program established foundational knowledge: Benchrif et al. [18] characterized air mass pathways, while subsequent studies quantified OC-to-POM and Ca²⁺-to-dust conversion factors crucial for applying the reconstructed mass balance method [15] and demonstrated the potential of integrating three complementary approaches—source-receptor modelling, chemical mass closure, and trajectory statistical methods—to gain a comprehensive understanding of PM composition, emission sources, and origin areas [23].

Despite these prior efforts, the Southwestern Mediterranean remains critically understudied, with major gaps in understanding how seasonal meteorology interacts with local and regional emissions to shape aerosol composition and, consequently, air quality in North African urban environments. This study is motivated by the need to fill this knowledge gap and provide actionable insights into the temporal dynamics of aerosol mass and chemistry in Tetouan. A comprehensive characterization of aerosol temporal variations is widely recognized as essential for identifying seasonal trends, episodic pollution events, and supporting the development of effective mitigation and adaptation strategies [24,25]. The main objective is to quantify the seasonal variations in PM_2.5 and PM₁₀ mass and chemical composition—focusing on carbonaceous species, water-soluble ions, and sugar tracers—while determining the relative contributions of local emissions and long-range transport using source-receptor modelling and trajectory analysis. This study provides a unique dataset and an integrative assessment of aerosol behaviour in a complex Mediterranean urban setting, delivering critical information for regional air quality management and advancing our understanding of aerosol processes in the Southwestern Mediterranean.

2. Materials and Methods

2.1. Aerosol Sampling, Analysis, and Quality Assurance/Quality Control

Continuous daily sampling of PM₁₀ and PM_2.5 (aerosols with aerodynamic diameters ≤10 µm and ≤2.5 µm, respectively) was conducted in Tetouan city, Northern Morocco, from May 2011 to April 2012. The sampling site (35.57° N, 5.36° W), classified as an urban background site based on the criteria outlined by Van Dingenen et al. [26], was situated approximately 1.5 km from the historic city centre (Figure 1). The urban environment of Tetouan city is influenced by a mixture of local emission sources. These include vehicular traffic emissions, marine aerosols from the nearby Mediterranean Sea, and the resuspension of mineral dust, especially under dry and windy conditions. Previous studies in the region have identified these sources as major contributors to PM_2.5 and PM₁₀, with distinct seasonal patterns and chemical signatures. Specifically, traffic-related emissions and biomass burning contribute significantly to carbonaceous aerosol levels, while aged sea salt and secondary inorganic aerosols reflect both local and regional atmospheric processes [18,23,27]. The sampling took place on the rooftop of the Artisan School, 15 m above ground level. PM was collected using two Teflon-coated aluminium cyclones (URG Corp., Chapel Hill, CA, USA) with aerodynamic cut points at 10 µm (model URG-2000-30ENB, URG Corp., Chapel Hill, NC, USA) and 2.5 µm (model URG-2000-30EH, URG Corp., Chapel Hill, NC, USA), operating at a flow rate of 1 m³ h⁻¹. Sampling durations were dynamically adjusted to ensure collection efficiency and cut-off reliability. Each PM fraction was collected on 47 mm diameter filters: Teflon filters (Zefluor, Pall Corp., Port Washington, NY, USA) for gravimetric, ion, and sugar analyses, and pre-baked quartz fibre filters (QM/A, Whatman Inc., Middlesex, UK) for carbonaceous fraction quantification. Quartz fibre filters were pre-baked at 400 °C for 48 h following the procedure of Guinot et al. [28], in order to minimize artifacts and remove potential residual organic matter while maintaining low blank values. Table 1 summarizes characteristics of the main datasets, analytical techniques, and quality assurance measures. A total of 89 samples were collected following a systematic schedule to ensure temporal representativeness rather than random sampling. Two sampling periods were performed each week, on weekdays and weekends. This design allowed the capture of both seasonal and weekly variations in aerosol composition. The number of samples collected per season is indicated in Table 2.

Before and after sampling, filters were conditioned in a desiccator (T ~25 °C, RH < 30%) for at least 24 h to minimize hygroscopic effects on weight measurements [29]. Weighing was performed using a high-precision microbalance (UMT3, Mettler Toledo Inc., Greifensee, Switzerland; sensitivity ±1 µg). Gravimetric mass uncertainty was typically around 20 µg, derived from repeated weighing of blank filters and accounting for balance precision, environmental variability, and handling errors [28,30]. This corresponds to average relative uncertainties of 2.0% for PM_2.5 and 3.8% for PM₁₀. Field blanks were collected in the same conditions as the samples. Following collection, the filters were stored in a cool and dark environment until analysis. OC and EC were quantified via coulometry (Ströhlein Coulomat 702C) using a two-step combustion method [31], with blank corrections applied. Measurement uncertainties were ~5%, with blank correction contributions averaging 2% for EC and 7% for OC, translating to total uncertainties of 4.5% (EC) and 7.6% (OC). For water-soluble inorganic ions (SO₄²⁻, NO₃⁻, Cl⁻, C₂O₄²⁻, Na⁺, NH₄⁺, K⁺, Mg²⁺, Ca²⁺), ion chromatography (IC) was performed using a Dionex^® (model DX-600, Dionex, Sunnyvale, CA, USA). Sugar compounds were analysed by IC coupled with Pulsed Amperometric Detection (IC/PAD). Filters were extracted in ultrapure water under mechanical agitation for 40 min. Field blanks were processed identically. Maximum uncertainty for major ion analysis was estimated at 0.050 µg m⁻³ (less than 1%). Analytical reproducibility for major sugar species was assessed via ten replicates, yielding relative standard deviations between 10% and 15%.

The non-sea-salt (nss) aerosol contributions for each ion were quantified by subtracting the estimated sea-salt fraction from the total measured concentration. As summarized in Table 1, the sea-salt fraction of Cl⁻, SO₄²⁻, Ca²⁺, Mg²⁺, and K⁺ was calculated using their standard ratios to sodium (Na⁺) in seawater following Seinfeld et al. [32].

Table 1. Summary of measured parameters and data characteristics.

Parameter Category	Measured Parameter	Method/Instrument	Units	Temporal Resolution	Relative Uncertainty	Quality Control Measures
Particulate Matter	PM2.5 and PM10 mass	Gravimetric/High-precision microbalance	µg m−3	24-h	2.0% (PM2.5), 3.8% (PM10)	Field blanks, replicate weighing
Carbonaceous Aerosols	OC, EC	Two-step combustion/Coulometry (Ströhlein 702C)		24-h	4.5% (EC), 7.6% (OC)	Glucose standard, blank correction
	SOC	EC tracer method/Estimated: SOC = OC − (OC-to-EC)pri × EC − POCnon-comb
	TCA	Estimated: TCA = 1.2 × OC + EC				Based on the OM/OC factor validated in Benchrif et al. [15]
Water-Soluble Ions	SO42−, NO3−, Cl−, C2O42−, Na+, NH4+, K+, Mg2+, Ca2+	Ion Chromatography (Dionex DX-600)		24-h	<1%	IC calibration, anion-cation balance
	Sea-salt contribution	Estimated: [ss-Cl−] = [Na+] × 1.8; [ss-SO42−] = [Na+] × 0.252; [ss-Ca2+] = [Na+] × 0.038; [ss-Mg2+] = [Na+] × 0.12; [ss-K+] = [Na+] × 0.036				The sea-salt fraction was calculated using their standard ratios to sodium (Na+) in seawater following Seinfeld et al. [32]
Sugar Compounds	Levoglucosan, arabitol, glucose	IC-PAD	ng m−3	24-h	10–15% RSD	Standard calibration, blank correction
Meteorological Data	Temperature, RH, wind speed, rainfall	Local meteorological station (Sania-Ramel, 35.58° N, 5.33° W)	°C, %, m s−1, mm	6-h
	Boundary Layer Height	HYSPLIT (GDAS)	m			Data from GDAS (available at https://webspace.clarkson.edu/projects/TraPSA/public_html/en/downloaddata.html, accessed on 1 March 2021)
Air Mass Trajectories	72-h back trajectories (CWT and Clustering)	HYSPLIT (GDAS)		6-h endpoints (trajectories)		Data from GDAS and trajectory convergence check

Table 1. Summary of measured parameters and data characteristics.

Parameter Category	Measured Parameter	Method/Instrument	Units	Temporal Resolution	Relative Uncertainty	Quality Control Measures
Particulate Matter	PM2.5 and PM10 mass	Gravimetric/High-precision microbalance	µg m−3	24-h	2.0% (PM2.5), 3.8% (PM10)	Field blanks, replicate weighing
Carbonaceous Aerosols	OC, EC	Two-step combustion/Coulometry (Ströhlein 702C)		24-h	4.5% (EC), 7.6% (OC)	Glucose standard, blank correction
	SOC	EC tracer method/Estimated: SOC = OC − (OC-to-EC)pri × EC − POCnon-comb
	TCA	Estimated: TCA = 1.2 × OC + EC				Based on the OM/OC factor validated in Benchrif et al. [15]
Water-Soluble Ions	SO42−, NO3−, Cl−, C2O42−, Na+, NH4+, K+, Mg2+, Ca2+	Ion Chromatography (Dionex DX-600)		24-h	<1%	IC calibration, anion-cation balance
	Sea-salt contribution	Estimated: [ss-Cl−] = [Na+] × 1.8; [ss-SO42−] = [Na+] × 0.252; [ss-Ca2+] = [Na+] × 0.038; [ss-Mg2+] = [Na+] × 0.12; [ss-K+] = [Na+] × 0.036				The sea-salt fraction was calculated using their standard ratios to sodium (Na+) in seawater following Seinfeld et al. [32]
Sugar Compounds	Levoglucosan, arabitol, glucose	IC-PAD	ng m−3	24-h	10–15% RSD	Standard calibration, blank correction
Meteorological Data	Temperature, RH, wind speed, rainfall	Local meteorological station (Sania-Ramel, 35.58° N, 5.33° W)	°C, %, m s−1, mm	6-h
	Boundary Layer Height	HYSPLIT (GDAS)	m			Data from GDAS (available at https://webspace.clarkson.edu/projects/TraPSA/public_html/en/downloaddata.html, accessed on 1 March 2021)
Air Mass Trajectories	72-h back trajectories (CWT and Clustering)	HYSPLIT (GDAS)		6-h endpoints (trajectories)		Data from GDAS and trajectory convergence check

Table 2. Descriptive statistics (mean ± standard deviation, [minimum–maximum]) of PM₁₀ and PM_2.5 concentrations (µg m⁻³), major PM chemical constituents (µg m⁻³), and 6-h meteorological parameters (T (temperature, °C), Rainfall (mm), WS (wind speed, m s⁻¹), RH (relative humidity, %), BLH (boundary layer height, m), and VC (ventilation coefficient, m² s⁻¹)) between mid-May 2011 to April 2012. Summary data are subdivided into four seasons over Tetouan city. n stands for the number of total datasets used. n.a. stands for not available data. For sugar compounds (ng m⁻³), the number of datasets used in summer and spring is 23 and 21, respectively.

Specie	Overall (n = 89)		Summer (n = 23)		Autumn (n = 19)		Winter (n = 25)		Spring (n = 22)
	PM10	PM2.5	PM10	PM2.5	PM10	PM2.5	PM10	PM2.5	PM10	PM2.5
PM	30.8 ± 9.7 [11.9–66.3]	18.0 ± 6.4 [4.2–41.8]	34.7 ± 9.9 [19.5–54.6]	17.8 ± 3.7 [11.1–23.6]	26.2 ± 6.3 [11.9–36.0]	16.7 ± 4.9 [6.6–27.0]	32.3 ± 10.4 [18.0–66.3]	21.0 ± 8.0 [7.9–41.8]	29.1 ± 9.4 [15.2–43.1]	15.8 ± 6.9 [4.2–27.1]
EC	3.0 ± 1.2 [1.1–7.0]	3.2 ± 1.6 [0.5–8.4]	3.4 ± 1.4 [1.4–6.2]	3.9 ± 1.9 [0.8–7.8]	3.2 ± 1.2 [1.2–5.3]	3.4 ± 1.1 [1.6–5.5]	3.2 ± 1.2 [1.5–7.0]	3.2 ± 1.6 [1.7–8.4]	2.3 ± 0.8 [1.1–4.8]	2.6 ± 1.4 [0.5–6.4]
OC	5.7 ± 2.4 [1.1–14.1]	5.4 ± 2.8 [0.1–18.2]	4.8 ± 1.2 [2.6–6.9]	3.9 ± 2.1 [0.1–11.1]	6.6 ± 1.9 [2.6–10.3]	5.8 ± 1.5 [2.8–9.3]	7.0 ± 3.2 [1.1–14.1]	7.7 ± 3.4 [2.1–18.2]	4.4 ± 1.1 [2.8–6.3]	3.9 ± 1.0 [2.2–5.7]
OC-to-EC	2.1 ± 1.2 [0.4–8.6]	2.1 ± 1.6 [0.2–10]	1.7 ± 1.0 [0.5–4.8]	1.2 ± 0.8 [0.2–3.4]	2.6 ± 1.8 [0.5–8.6]	1.9 ± 1.0 [0.7–4.5]	2.3 ± 0.9 [0.4–4.8]	2.7 ± 1.0 [0.3–4.6]	2.1 ± 0.7 [1.0–3.5]	2.4 ± 2.5 [0.5–10]
TCA	9.9 ± 3.4 [4.1–23.8]	9.6 ± 3.7 [1.0–28.6]	9.2 ± 1.8 [5.5–13.0]	8.0 ± 2.6 [1.0–12.3]	11.1 ± 2.1 [6.7–14.6]	10.3 ± 1.6 [7.3–14.5]	11.6 ± 4.8 [4.1–23.8]	12.5 ± 4.7 [6.6–28.6]	7.5 ± 1.8 [5.2–11.9]	7.1 ± 1.1 [5.1–8.9]
TCA-to-PM	0.34 ± 0.13 [0.16–0.83]	0.52 ± 0.16 [0.05–0.96]	0.28 ± 0.07 [0.16–0.42]	0.45 ± 0.15 [0.05–0.77]	0.45 ± 0.15 [0.22–0.80]	0.59 ± 0.15 [0.32–0.96]	0.37 ± 0.14 [0.18–0.83]	0.59 ± 0.13 [0.38–0.85]	0.28 ± 0.09 [0.17–0.51]	0.44 ± 0.17 [0.25–0.81]
POC	1.5 ± 0.6 [0.6–3.6]	0.4 ± 0.2 [0.1–0.9]	1.7 ± 0.7 [0.7–3.1]	0.4 ± 0.2 [0.1–0.9]	1.6 ± 0.6 [0.6–2.6]	0.4 ± 0.1 [0.2–0.6]	1.6 ± 0.6 [0.7–3.5]	0.4 ± 0.2 [0.2–0.9]	1.1 ± 0.4 [0.6–2.4]	0.3 ± 0.2 [0.1–0.7]
SOC	4.3 ± 2.2 [0.8–12.5]	5.0 ± 2.8 [0.1–17.4]	3.2 ± 1.4 [0.8–6.2]	3.5 ± 2.0 [0.1–10.2]	5.3 ± 1.8 [2.7–9.7]	5.4 ± 1.6 [2.4–9.0]	5.7 ± 2.6 [1.4–12.5]	7.3 ± 3.4 [1.2–17.4]	3.2 ± 1.1 [1.8–5.3]	3.6 ± 1.1 [1.9–5.6]
SOC-to-OC	0.7 ± 0.1 [0.3–0.9]	0.9 ± 0.1 [0.3–1.0]	0.6 ± 0.2 [0.3–0.9]	0.9 ± 0.1 [0.3–1.0]	0.7 ± 0.1 [0.6–0.9]	0.9 ± 0.0 [0.8–1.0]	0.8 ± 0.1 [0.5–0.9]	0.9 ± 0.1 [0.6–1.0]	0.7 ± 0.1 [0.5–0.9]	0.9 ± 0.1 [0.8–1.0]
Na+	1.4 ± 0.8 [0.3–3.9]	0.5 ± 0.3 [0.1–1.7]	1.9 ± 0.9 [0.4–3.9]	0.7 ± 0.3 [0.2–1.7]	1.4 ± 0.6 [0.3–2.5]	0.4 ± 0.3 [0.1–0.9]	1.0 ± 0.6 [0.5–2.5]	0.4 ± 0.2 [0.2–0.8]	1.4 ± 0.8 [0.3–3.6]	0.5 ± 0.3 [0.1–1.2]
NH4+	0.8 ± 0.5 [0.1–2.8]	1.1 ± 0.7 [0.0–3.0]	0.9 ± 0.4 [0.3–1.7]	1.4 ± 0.6 [0.7–2.7]	0.5 ± 0.5 [0.1–1.9]	0.6 ± 0.7 [0.0–3.0]	1.0 ± 0.7 [0.2–2.8]	1.2 ± 0.8 [0.3–3.0]	0.7 ± 0.5 [0.1–1.9]	1.0 ± 0.6 [0.1–2.1]
K+	0.5 ± 0.3 [0.0–2.0]	0.4 ± 0.3 [0.0–1.7]	0.3 ± 0.2 [0.1–0.7]	0.3 ± 0.1 [0.2–0.7]	0.5 ± 0.2 [0.1–0.9]	0.4 ± 0.2 [0.1–0.8]	0.8 ± 0.4 [0.3–2.0]	0.7 ± 0.4 [0.3–1.7]	0.4 ± 0.2 [0.0–1.1]	0.3 ± 0.2 [0.0–0.9]
Mg2+	0.2 ± 0.1 [0.1–0.6]	0.1 ± 0.1 [0.0–0.3]	0.3 ± 0.1 [0.1–0.6]	0.2 ± 0.1 [0.1–0.3]	0.2 ± 0.1 [0.1–0.4]	0.1 ± 0.0 [0.0–0.1]	0.2 ± 0.1 [0.1–0.3]	0.0 [0.0–0.1]	0.3 ± 0.1 [0.1–0.5]	0.1 ± 0.1 [0.0–0.2]
Ca2+	2.8 ± 0.9 [1.0–5.3]	0.4 ± 0.3 [0.1–2.5]	2.6 ± 0.6 [1.2–4.1]	0.6 ± 0.5 [0.2–2.5]	2.6 ± 0.8 [1.7–4.1]	0.3 ± 0.2 [0.1–0.6]	3.3 ± 0.9 [1.8–5.3]	0.3 ± 0.1 [0.2–0.8]	2.6 ± 0.9 [1.0–4.3]	0.4 ± 0.2 [0.1–0.9]
Cl−	1.5 ± 1.0 [0.0–4.9]	0.4 ± 0.4 [0.0–2.1]	1.4 ± 1.2 [0.0–4.9]	0.2 ± 0.3 [0.0–0.9]	1.5 ± 0.7 [0.3–3.0]	0.3 ± 0.3 [0.0–0.8]	1.6 ± 1.0 [0.4–3.9]	0.7 ± 0.5 [0.1–2.1]	1.4 ± 1.2 [0.2–4.1]	0.3 ± 0.3 [0.0–1.2]
NO3−	2.5 ± 1.3 [0.2–6.6]	1.0 ± 0.8 [0.1–4.0]	2.5 ± 1.3 [0.2–5.9]	0.8 ± 0.5 [0.3–3.0]	2.1 ± 1.0 [0.7–4.4]	0.7 ± 0.4 [0.1–1.9]	3.0 ± 1.5 [0.8–6.6]	1.5 ± 1.0 [0.4–3.9]	2.2 ± 1.0 [0.5–4.1]	1.0 ± 0.8 [0.3–4.0]
SO42−	3.6 ± 1.9 [0.7–9.0]	3.0 ± 1.7 [0.2–6.6]	5.0 ± 2.0 [2.5–9.0]	4.3 ± 1.4 [2.2–6.6]	2.9 ± 2.0 [0.7–8.8]	2.0 ± 1.6 [0.2–6.5]	3.0 ± 1.1 [1.2–5.4]	2.6 ± 1.2 [0.9–5.3]	3.6 ± 1.9 [0.9–6.9]	2.9 ± 1.7 [0.4–5.8]
C2O42−	0.1 ± 0.1 [0.0–0.4]	0.1 ± 0.1 [0.0–0.3]	0.2 ± 0.1 [0.0–0.4]	0.0 ± 0.1 [0.0–0.3]	0.1 ± 0.1 [0.0–0.2]	0.1 ± 0.1 [0.0–0.2]	0.1 ± 0.1 [0.1–0.2]	0.1 ± 0.1 [0.0–0.2]	0.1 ± 0.1 [0.0–0.3]	0.1 ± 0.1 [0.0–0.2]
WSI	13.5 ± 4.6 [4.4–23.6]	7.0 ± 3.3 [0.6–17.6]	15.1 ± 4.9 [7.5–23.6]	8.5 ± 2.8 [5.2–17.6]	11.8 ± 4.3 [4.4–20.8]	4.7 ± 2.7 [0.6–11.0]	14.0 ± 4.4 [6.2–22.3]	7.7 ± 3.5 [3.0–14.8]	12.6 ± 4.2 [5.8–20.8]	6.6 ± 3.0 [1.6–12.0]
SIA	6.6 ± 3.1 [1.5–13.9]	5.0 ± 2.7 [0.3–11.6]	8.0 ± 3.1 [3.9–13.8]	6.4 ± 2.1 [3.3–11.0]	5.2 ± 3.0 [1.5–12.4]	3.2 ± 2.4 [0.3–9.9]	6.7 ± 2.9 [2.4–13.9]	5.2 ± 2.8 [1.5–11.6]	6.2 ± 3.0 [1.7–11.0]	4.9 ± 2.5 [0.8–9.1]
Levoglucosan	21.7 ± 16.9 [1.5–84.0]	10.8 ± 11.5 [0.5–49.2]	17.2 ± 8.8 [4.2–35.4]	5.5 ± 6.1 [0.5–26.4]	n.a.	n.a.	n.a.	n.a.	26.7 ± 21.9 [1.5–84.0]	16.6 ± 13.3 [1.0–49.2]
Arabitol	1.93 ± 2.58 [0.0–11.70]	0.23 ± 0.26 [0.0–1.26]	0.67 ± 1.25 [0.0–6.24]	0.23 ± 0.17 [0.0–0.48]	n.a.	n.a.	n.a.	n.a.	3.30 ± 2.97 [0.0–11.7]	0.22 ± 0.33 [0.0–1.26]
Glucose	10 ± 15 [0.1–67.7]	2.2 ± 2.4 [0.1–9.6]	6.2 ± 8.8 [0.3–41.2]	2.3 ± 3.0 [0.2–9.6]	n.a.	n.a.	n.a.	n.a.	14.1 ± 18.9 [0.1–67.7]	2.0 ± 1.5 [0.1–5.2]
T	19.2 [9.9–27.9]		25.1 [22.3–27.9]		18.0 [13.6–22.5]		13.7 [9.9–16.4]		20.4 [14.7–27.0]
Rainfall	0.9 [0.0–24.1]		0.1 [0.0–0.8]		1.3 [0.0–11.3]		0.7 [0.0–6.3]		1.7 [0.0–24.1]
WS	4.4 [1.6–9.8]		4.2 [2.7–6.8]		4.0 [2.4–5.4]		4.6 [2.9–9.1]		4.7 [1.6–9.8]
RH	67 [36–90]		65 [46–80]		70 [55–81]		66 [48–86]		68 [36–90]
BLH	461 [207–1025]		400 [207–605]		477 [283–725]		480 [280–846]		490 [221–1024]
VC	2146 [431–10,051]		1724 [558–3677]		1922 [953–3748]		2373 [864–7729]		2523 [431–10,051]

Table 2. Descriptive statistics (mean ± standard deviation, [minimum–maximum]) of PM₁₀ and PM_2.5 concentrations (µg m⁻³), major PM chemical constituents (µg m⁻³), and 6-h meteorological parameters (T (temperature, °C), Rainfall (mm), WS (wind speed, m s⁻¹), RH (relative humidity, %), BLH (boundary layer height, m), and VC (ventilation coefficient, m² s⁻¹)) between mid-May 2011 to April 2012. Summary data are subdivided into four seasons over Tetouan city. n stands for the number of total datasets used. n.a. stands for not available data. For sugar compounds (ng m⁻³), the number of datasets used in summer and spring is 23 and 21, respectively.

Specie	Overall (n = 89)		Summer (n = 23)		Autumn (n = 19)		Winter (n = 25)		Spring (n = 22)
	PM10	PM2.5	PM10	PM2.5	PM10	PM2.5	PM10	PM2.5	PM10	PM2.5
PM	30.8 ± 9.7 [11.9–66.3]	18.0 ± 6.4 [4.2–41.8]	34.7 ± 9.9 [19.5–54.6]	17.8 ± 3.7 [11.1–23.6]	26.2 ± 6.3 [11.9–36.0]	16.7 ± 4.9 [6.6–27.0]	32.3 ± 10.4 [18.0–66.3]	21.0 ± 8.0 [7.9–41.8]	29.1 ± 9.4 [15.2–43.1]	15.8 ± 6.9 [4.2–27.1]
EC	3.0 ± 1.2 [1.1–7.0]	3.2 ± 1.6 [0.5–8.4]	3.4 ± 1.4 [1.4–6.2]	3.9 ± 1.9 [0.8–7.8]	3.2 ± 1.2 [1.2–5.3]	3.4 ± 1.1 [1.6–5.5]	3.2 ± 1.2 [1.5–7.0]	3.2 ± 1.6 [1.7–8.4]	2.3 ± 0.8 [1.1–4.8]	2.6 ± 1.4 [0.5–6.4]
OC	5.7 ± 2.4 [1.1–14.1]	5.4 ± 2.8 [0.1–18.2]	4.8 ± 1.2 [2.6–6.9]	3.9 ± 2.1 [0.1–11.1]	6.6 ± 1.9 [2.6–10.3]	5.8 ± 1.5 [2.8–9.3]	7.0 ± 3.2 [1.1–14.1]	7.7 ± 3.4 [2.1–18.2]	4.4 ± 1.1 [2.8–6.3]	3.9 ± 1.0 [2.2–5.7]
OC-to-EC	2.1 ± 1.2 [0.4–8.6]	2.1 ± 1.6 [0.2–10]	1.7 ± 1.0 [0.5–4.8]	1.2 ± 0.8 [0.2–3.4]	2.6 ± 1.8 [0.5–8.6]	1.9 ± 1.0 [0.7–4.5]	2.3 ± 0.9 [0.4–4.8]	2.7 ± 1.0 [0.3–4.6]	2.1 ± 0.7 [1.0–3.5]	2.4 ± 2.5 [0.5–10]
TCA	9.9 ± 3.4 [4.1–23.8]	9.6 ± 3.7 [1.0–28.6]	9.2 ± 1.8 [5.5–13.0]	8.0 ± 2.6 [1.0–12.3]	11.1 ± 2.1 [6.7–14.6]	10.3 ± 1.6 [7.3–14.5]	11.6 ± 4.8 [4.1–23.8]	12.5 ± 4.7 [6.6–28.6]	7.5 ± 1.8 [5.2–11.9]	7.1 ± 1.1 [5.1–8.9]
TCA-to-PM	0.34 ± 0.13 [0.16–0.83]	0.52 ± 0.16 [0.05–0.96]	0.28 ± 0.07 [0.16–0.42]	0.45 ± 0.15 [0.05–0.77]	0.45 ± 0.15 [0.22–0.80]	0.59 ± 0.15 [0.32–0.96]	0.37 ± 0.14 [0.18–0.83]	0.59 ± 0.13 [0.38–0.85]	0.28 ± 0.09 [0.17–0.51]	0.44 ± 0.17 [0.25–0.81]
POC	1.5 ± 0.6 [0.6–3.6]	0.4 ± 0.2 [0.1–0.9]	1.7 ± 0.7 [0.7–3.1]	0.4 ± 0.2 [0.1–0.9]	1.6 ± 0.6 [0.6–2.6]	0.4 ± 0.1 [0.2–0.6]	1.6 ± 0.6 [0.7–3.5]	0.4 ± 0.2 [0.2–0.9]	1.1 ± 0.4 [0.6–2.4]	0.3 ± 0.2 [0.1–0.7]
SOC	4.3 ± 2.2 [0.8–12.5]	5.0 ± 2.8 [0.1–17.4]	3.2 ± 1.4 [0.8–6.2]	3.5 ± 2.0 [0.1–10.2]	5.3 ± 1.8 [2.7–9.7]	5.4 ± 1.6 [2.4–9.0]	5.7 ± 2.6 [1.4–12.5]	7.3 ± 3.4 [1.2–17.4]	3.2 ± 1.1 [1.8–5.3]	3.6 ± 1.1 [1.9–5.6]
SOC-to-OC	0.7 ± 0.1 [0.3–0.9]	0.9 ± 0.1 [0.3–1.0]	0.6 ± 0.2 [0.3–0.9]	0.9 ± 0.1 [0.3–1.0]	0.7 ± 0.1 [0.6–0.9]	0.9 ± 0.0 [0.8–1.0]	0.8 ± 0.1 [0.5–0.9]	0.9 ± 0.1 [0.6–1.0]	0.7 ± 0.1 [0.5–0.9]	0.9 ± 0.1 [0.8–1.0]
Na+	1.4 ± 0.8 [0.3–3.9]	0.5 ± 0.3 [0.1–1.7]	1.9 ± 0.9 [0.4–3.9]	0.7 ± 0.3 [0.2–1.7]	1.4 ± 0.6 [0.3–2.5]	0.4 ± 0.3 [0.1–0.9]	1.0 ± 0.6 [0.5–2.5]	0.4 ± 0.2 [0.2–0.8]	1.4 ± 0.8 [0.3–3.6]	0.5 ± 0.3 [0.1–1.2]
NH4+	0.8 ± 0.5 [0.1–2.8]	1.1 ± 0.7 [0.0–3.0]	0.9 ± 0.4 [0.3–1.7]	1.4 ± 0.6 [0.7–2.7]	0.5 ± 0.5 [0.1–1.9]	0.6 ± 0.7 [0.0–3.0]	1.0 ± 0.7 [0.2–2.8]	1.2 ± 0.8 [0.3–3.0]	0.7 ± 0.5 [0.1–1.9]	1.0 ± 0.6 [0.1–2.1]
K+	0.5 ± 0.3 [0.0–2.0]	0.4 ± 0.3 [0.0–1.7]	0.3 ± 0.2 [0.1–0.7]	0.3 ± 0.1 [0.2–0.7]	0.5 ± 0.2 [0.1–0.9]	0.4 ± 0.2 [0.1–0.8]	0.8 ± 0.4 [0.3–2.0]	0.7 ± 0.4 [0.3–1.7]	0.4 ± 0.2 [0.0–1.1]	0.3 ± 0.2 [0.0–0.9]
Mg2+	0.2 ± 0.1 [0.1–0.6]	0.1 ± 0.1 [0.0–0.3]	0.3 ± 0.1 [0.1–0.6]	0.2 ± 0.1 [0.1–0.3]	0.2 ± 0.1 [0.1–0.4]	0.1 ± 0.0 [0.0–0.1]	0.2 ± 0.1 [0.1–0.3]	0.0 [0.0–0.1]	0.3 ± 0.1 [0.1–0.5]	0.1 ± 0.1 [0.0–0.2]
Ca2+	2.8 ± 0.9 [1.0–5.3]	0.4 ± 0.3 [0.1–2.5]	2.6 ± 0.6 [1.2–4.1]	0.6 ± 0.5 [0.2–2.5]	2.6 ± 0.8 [1.7–4.1]	0.3 ± 0.2 [0.1–0.6]	3.3 ± 0.9 [1.8–5.3]	0.3 ± 0.1 [0.2–0.8]	2.6 ± 0.9 [1.0–4.3]	0.4 ± 0.2 [0.1–0.9]
Cl−	1.5 ± 1.0 [0.0–4.9]	0.4 ± 0.4 [0.0–2.1]	1.4 ± 1.2 [0.0–4.9]	0.2 ± 0.3 [0.0–0.9]	1.5 ± 0.7 [0.3–3.0]	0.3 ± 0.3 [0.0–0.8]	1.6 ± 1.0 [0.4–3.9]	0.7 ± 0.5 [0.1–2.1]	1.4 ± 1.2 [0.2–4.1]	0.3 ± 0.3 [0.0–1.2]
NO3−	2.5 ± 1.3 [0.2–6.6]	1.0 ± 0.8 [0.1–4.0]	2.5 ± 1.3 [0.2–5.9]	0.8 ± 0.5 [0.3–3.0]	2.1 ± 1.0 [0.7–4.4]	0.7 ± 0.4 [0.1–1.9]	3.0 ± 1.5 [0.8–6.6]	1.5 ± 1.0 [0.4–3.9]	2.2 ± 1.0 [0.5–4.1]	1.0 ± 0.8 [0.3–4.0]
SO42−	3.6 ± 1.9 [0.7–9.0]	3.0 ± 1.7 [0.2–6.6]	5.0 ± 2.0 [2.5–9.0]	4.3 ± 1.4 [2.2–6.6]	2.9 ± 2.0 [0.7–8.8]	2.0 ± 1.6 [0.2–6.5]	3.0 ± 1.1 [1.2–5.4]	2.6 ± 1.2 [0.9–5.3]	3.6 ± 1.9 [0.9–6.9]	2.9 ± 1.7 [0.4–5.8]
C2O42−	0.1 ± 0.1 [0.0–0.4]	0.1 ± 0.1 [0.0–0.3]	0.2 ± 0.1 [0.0–0.4]	0.0 ± 0.1 [0.0–0.3]	0.1 ± 0.1 [0.0–0.2]	0.1 ± 0.1 [0.0–0.2]	0.1 ± 0.1 [0.1–0.2]	0.1 ± 0.1 [0.0–0.2]	0.1 ± 0.1 [0.0–0.3]	0.1 ± 0.1 [0.0–0.2]
WSI	13.5 ± 4.6 [4.4–23.6]	7.0 ± 3.3 [0.6–17.6]	15.1 ± 4.9 [7.5–23.6]	8.5 ± 2.8 [5.2–17.6]	11.8 ± 4.3 [4.4–20.8]	4.7 ± 2.7 [0.6–11.0]	14.0 ± 4.4 [6.2–22.3]	7.7 ± 3.5 [3.0–14.8]	12.6 ± 4.2 [5.8–20.8]	6.6 ± 3.0 [1.6–12.0]
SIA	6.6 ± 3.1 [1.5–13.9]	5.0 ± 2.7 [0.3–11.6]	8.0 ± 3.1 [3.9–13.8]	6.4 ± 2.1 [3.3–11.0]	5.2 ± 3.0 [1.5–12.4]	3.2 ± 2.4 [0.3–9.9]	6.7 ± 2.9 [2.4–13.9]	5.2 ± 2.8 [1.5–11.6]	6.2 ± 3.0 [1.7–11.0]	4.9 ± 2.5 [0.8–9.1]
Levoglucosan	21.7 ± 16.9 [1.5–84.0]	10.8 ± 11.5 [0.5–49.2]	17.2 ± 8.8 [4.2–35.4]	5.5 ± 6.1 [0.5–26.4]	n.a.	n.a.	n.a.	n.a.	26.7 ± 21.9 [1.5–84.0]	16.6 ± 13.3 [1.0–49.2]
Arabitol	1.93 ± 2.58 [0.0–11.70]	0.23 ± 0.26 [0.0–1.26]	0.67 ± 1.25 [0.0–6.24]	0.23 ± 0.17 [0.0–0.48]	n.a.	n.a.	n.a.	n.a.	3.30 ± 2.97 [0.0–11.7]	0.22 ± 0.33 [0.0–1.26]
Glucose	10 ± 15 [0.1–67.7]	2.2 ± 2.4 [0.1–9.6]	6.2 ± 8.8 [0.3–41.2]	2.3 ± 3.0 [0.2–9.6]	n.a.	n.a.	n.a.	n.a.	14.1 ± 18.9 [0.1–67.7]	2.0 ± 1.5 [0.1–5.2]
T	19.2 [9.9–27.9]		25.1 [22.3–27.9]		18.0 [13.6–22.5]		13.7 [9.9–16.4]		20.4 [14.7–27.0]
Rainfall	0.9 [0.0–24.1]		0.1 [0.0–0.8]		1.3 [0.0–11.3]		0.7 [0.0–6.3]		1.7 [0.0–24.1]
WS	4.4 [1.6–9.8]		4.2 [2.7–6.8]		4.0 [2.4–5.4]		4.6 [2.9–9.1]		4.7 [1.6–9.8]
RH	67 [36–90]		65 [46–80]		70 [55–81]		66 [48–86]		68 [36–90]
BLH	461 [207–1025]		400 [207–605]		477 [283–725]		480 [280–846]		490 [221–1024]
VC	2146 [431–10,051]		1724 [558–3677]		1922 [953–3748]		2373 [864–7729]		2523 [431–10,051]

2.2. Estimation of Secondary Organic Aerosols

The secondary organic carbon (SOC) can be estimated using elemental carbon (EC) as the tracer for primary emitted organic carbon (POC) [33,34]. SOC can be considered as a secondary OC formed during atmospheric aging. OC in this study is separated into POC and SOC counterparts:

OC = (POC_comb + POC_non-comb) + SOC

(1)

where POC_comb is the primary OC from combustion, and POC_non-comb is the primary OC concentration emitted from processes not involving combustion activities. The magnitude of this non-combustion primary OC (POC_non-comb) would be reflected in the intercept in the OC versus EC linear regression.

Since both EC and POC_comb are mostly emitted by combustion processes and exhibit a strong correlation [35], their association can be expressed using the following equation:

POC_comb = (OC-to-EC)_pri × EC

(2)

where (OC-to-EC)_pri is qualified as the primary OC-to-EC ratio in freshly emitted combustion aerosols [36].

Combining equations (Equations (1) and (2)), the SOC estimation is calculated as follows:

SOC = OC − (OC-to-EC)_pri × EC − POC_non-comb

(3)

The key of the EC tracer method is to isolate measurement periods during which a high primary contribution to OC is expected and calculate a representative value of the (OC-to-EC)_pri ratio [37]. This method assumes that (i) (OC-to-EC)_pri is constant and that there are periods during which POC has a dominant contribution to OC; (ii) a low OC-to-EC ratio is indicative of a high POC contribution to the total OC. Salma et al. [38] reported that this ratio is specific to site-, source-, and aerosol-type, and highly sensitive to the method used for its determination. Several studies base their estimation of (OC-to-EC)_pri on the minimum OC-to-EC ratio generally disregarding the intercept POC_non-comb in Equation (3) [39], estimation using the lower percentiles (e.g., 5–20%) of OC-to-EC [40], the minimum r-squared method (MRS) [41], as well as different regression techniques.

In this study, the (OC-to-EC)_pri was determined by the MRS method using a user-friendly computer program [42,43] in Igor Pro (WaveMetrics, Inc., Lake Oswego, OR, USA). Our previous studies [15,23] argued that the substantial proportion of OC in Tetouan may originate from traffic-related activities, although we emphasized the probable influence of other combustion sources such as biomass burning, waste burning, and forest fires. Thus, for SOC estimation in this study, we assumed that POC_non-comb is zeroed. Figure S1 (in the Supplementary Materials) illustrates the (OC-to-EC)_pri determination based on the MRS method in PM_2.5 and PM₁₀ fractions. This analysis is based on one-year data of OC and EC measurements conducted at an urban site in Tetouan.

In the collected PM, the total carbonaceous component (TCA) represents the combined concentration of organic matter (OM) and elemental carbon. Since OM is typically estimated from organic carbon measurements, a conversion factor is applied. Benchrif et al. [15] suggested a factor of 1.2 for Tetouan urban environments to convert OC to OM. Consequently, the calculation for TCA is as follows:

TCA = 1.2 × OC + EC

(4)

2.3. Backward Trajectory and Concentration-Weighted Trajectory Analysis

To investigate the long-range transport of aerosols to the receptor site from potential sources, 72-h backward air mass trajectories were calculated at a height of 500 m above ground level (AGL) for the entire sampling period. These trajectories were generated using the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model, developed by the National Oceanic and Atmospheric Administration (NOAA) Air Resources Laboratory (ARL), with input from the Global Data Assimilation System (GDAS) meteorological dataset [44]. The potential source regions influencing the PM chemical composition at the receptor site were identified through Concentration-Weighted Trajectory (CWT) analysis [45]. In this method, the study area surrounding the receptor site was divided into an i × j gridded array, where each grid cell (ij) was assigned a weighted concentration based on the residence time of trajectories passing through it. The CWT value ($C_{i,j}$) for each grid cell was calculated as follows [46]:

$$C_{i,j}={\displaystyle \frac{1}{{\sum }_{l=1}^M{\tau }_{i,j,l}}}{\sum }_{l=1}^M{C}_l{\tau }_{i,j,l}$$

(5)

where $C_{i,j}$ is the mean weighted concentration in the ijth grid, ${\tau }_{i,j,l}$ is the total number of endpoints in the ijth grid cell of trajectory l, $C_l$ is the measured PM concentration at the receptor site upon arrival of trajectory l, and M is termed the total number of trajectories.

2.4. Statistical Analyses

The Shapiro–Wilk test was conducted to assess the normality of concentration data for all species. Since the data did not follow a normal distribution, a non-parametric Spearman correlation test was employed to determine the correlation coefficients between different species. Three significance levels were considered: p-value < 0.1 (indicating 90% confidence level), p-value < 0.05 (95% confidence level), and p-value < 0.01 (99% confidence level). The statistical analyses were conducted using the R programming language [47] and its associated libraries, such as openair [48], ggplot2 [49], lubridate [50], and dplyr [51].

3. Results and Discussion

3.1. Temporal Variation in PM₁₀ and PM_2.5 Mass Concentrations

Table 2 summarizes the descriptive statistics for the chemical composition of PM_2.5 and PM₁₀ fractions. Regarding the national standard of 50 µg m⁻³, 90.4% of the daily PM₁₀ values remained compliant, with the 90.4th percentile at 42 µg m⁻³. For PM_2.5, the annual mean concentration (18 µg m⁻³) remained well below the national annual limit of 35 µg m⁻³, and no daily exceedances of the daily 45 µg m⁻³ threshold were recorded. PM₁₀ exceeded the daily limit on only five occasions, reflecting generally moderate air quality in the study area.

A one-way ANOVA test revealed statistically significant seasonal differences (p < 0.05) in the mass concentrations of both PM fractions. As illustrated in Figure 2 and Table 2, autumn showed the lowest average PM₁₀ concentration (26 µg m⁻³), while winter recorded the highest concentration, with a maximum value reaching 66 µg m⁻³. Summer (35 µg m⁻³, on average) and spring (29 µg m⁻³) exhibited intermediate values that fell between the autumn and winter extremes. PM_2.5 followed a similar seasonal trend, peaking in winter (21 µg m⁻³) and reaching a minimum in spring (16 µg m⁻³). The winter increase in PM concentrations is likely linked to the confluence of factors. First, reduced precipitation and lower temperature lead to the formation of secondary aerosols through enhanced gas-to-particle conversion [9,52]. Second, the prevalence of a shallower boundary layer (280–846 m) restricts vertical dispersion, trapping particles closer to ground level. Additionally, the increase in local combustion-related emissions, such as fossil fuel use and specific biomass burning practices [23,53].

To further investigate the size dependency of aerosols, the PM_2.5-to-PM₁₀ ratio was analysed. This ratio has been established as a valuable tool for characterizing local atmospheric processes [10]. It can reveal the dominant size fraction of aerosols in a specific region [54,55] and provide insights into the origins of PM pollution, differentiating between anthropogenic and natural sources [56]. For instance, smaller PM_2.5-to-PM₁₀ ratios suggest a greater contribution from natural processes, while higher ratios indicate a significant presence of fine aerosols originating from anthropogenic activities [10,57]. In Tetouan, the PM_2.5-to-PM₁₀ ratio exhibited significant variability, ranging from 0.23 to 0.92 with a mean value of 0.59, indicating a substantial distribution of particle mass within the fine fraction. A strong correlation (r = 0.77, p-value < 2.2 × 10⁻¹⁶) between PM₁₀ and PM_2.5 supports the hypothesis of shared emission sources [53,58]. Seasonal trends in the ratio were significant (p < 0.001), with higher values (~0.63) during colder seasons (winter and autumn) and lower ratios (~0.55) characterized the warm period (spring and summer). The average PM_2.5-to-PM₁₀ ratio in the cold period was approximately 1.2 times higher compared to the warm period. This seasonal variability is likely attributed to increased fine particle contributions resulting from prevailing meteorological conditions and enhanced secondary aerosol formation. The definitions of warm and cold periods were based on temperature profiles presented in Figure S2.

Comparative analysis with other Mediterranean locations provides further context (Table S1). The average PM_2.5 concentrations in Tetouan (18.0 ± 6.4 µg m⁻³; range: 4.2–41.8 µg m⁻³) align with values reported for cities in France, Lebanon, Malta, and Italy (15–20 µg m⁻³) [59,60,61]. However, it is higher than levels reported in rural areas, such as Montseny, Spain (5.55 µg m⁻³) [62] and Bou-Ismail, Algeria (12.3 µg m⁻³) [16]. Similarly, the average PM₁₀ concentration (30.8 ± 9.7 µg m⁻³; range: 11.9–66.3 µg m⁻³) is comparable to urban levels in Granada and Elche, Spain (~32 µg m⁻³) [63,64]. The PM_2.5-to-PM₁₀ ratio in Tetouan (0.59, on average) is consistent with urban Mediterranean environments: 0.56 in Athens, Greece [65], 0.57 in Beirut, Lebanon [66], and 0.5 in Nador, Morocco [67].

3.2. Concentration Distribution of Water-Soluble Inorganic Ions in PM₁₀ and PM_2.5 Aerosols

Water-soluble inorganic ions (WSIs) dominate aerosol toxicity and radiative effects [68,69]. This section aims to discuss their composition in Tetouan, a Mediterranean coastal city influenced by local emissions and aged marine air masses, to resolve seasonal source-receptor relationships. The monthly and seasonal variations in the concentrations of major ionic components are represented in Figure 3 and Figure S3, respectively. Statistical summaries of these components are presented in Table 2.

The majority of WSIs in PM₁₀ showed significant seasonal variability, with the highest concentrations generally occurring during the warm season, except for NO₃⁻, Cl⁻, Ca²⁺, and K⁺, which peaked in winter. SO₄²⁻ exhibited the highest annual average concentration among all nine selected WSIs (3.6 µg m⁻³), particularly in summer, reflecting enhanced photochemical oxidation of SO₂. Nitrate (2.5 µg m⁻³) and Ca²⁺ (2.8 µg m⁻³) followed, the former linked to heterogeneous reactions involving NOₓ, and the latter mainly originating from construction dust and road traffic resuspension [23,70,71]. The sampling site is influenced by Mediterranean aerosols [15,18], where aging of sea salt particles via acid displacement (e.g., with HNO₃) could explain particulate nitrate formation. Since seawater-derived nitrate is negligible [72], NO₃⁻ here primarily stems from gas-to-particle conversion of anthropogenic HNO₃. Ammonium (NH₄⁺), non-sea-salt sulfate (nss-SO₄²⁻), and nitrate—the key secondary inorganic aerosols (SIAs)—showed strong seasonal trends, with highest concentrations in summer (Figure S3). These are consistent with gas-to-particle conversion mechanisms of their corresponding precursors, NO_x, SO₂, and NH₃, during warm and stagnant atmospheric conditions. On average, SIAs accounted for ~50% of the total WSIs in PM₁₀, with nss-SO₄²⁻ the major component representing 25% of this fraction (accounting for 3.3 µg m⁻³).

In the PM_2.5 fraction, the total WSI concentration averaged 7.0 µg m⁻³, accounting for approximately 39% of the total PM_2.5 mass. Seasonal analysis revealed the highest levels in summer (8.2 µg m⁻³), likely driven by persistent thermal inversions, a shallow boundary layer (~375 m), and low wind speeds (~4.0 m s⁻¹), all of which favour the pollutant accumulation. Additionally, during the summer, the prevailing air masses in the region predominantly originate from the north, particularly from the Mediterranean Sea and the Atlantic Ocean, potentially introducing long-range transported aerosols [18]. A secondary peak was observed in winter (6.4 µg m⁻³), attributed to lower average temperatures (13.2 °C in January) and elevated relative humidity (70%), which enhance gas-to-particle partitioning processes, especially for nitrate and ammonium species [73]. Furthermore, anti-cyclonic pollution episodes during winter, indicated by elevated nitrate concentrations reaching 1.8 µg m⁻³, support the occurrence of enhanced gas-to-particle conversion under low-temperature and stagnant atmospheric conditions [13].

SIA species were dominant in PM_2.5, accounting for ~72% of the total WSI mass. These ions are mainly formed via physical/chemical reactions, such as gas-to-particle conversion through homogeneous nucleation, condensation, and coagulation processes [74,75]. It is further suggested that the high concentrations of secondary inorganic species in PM_2.5 may be attributed to significant contributions from long-range transported aerosols, particularly from anthropogenic and marine sources across the Mediterranean Basin. However, previous findings [15] also point to a dual origin of the fine aerosol fraction, as regional transport and local emissions. Seasonally, the SIA concentrations peaked in summer (6.4 µg m⁻³), followed by winter (5.2 µg m⁻³), spring (4.9 µg m⁻³), and autumn (3.2 µg m⁻³), contributing 75%, 68%, 74%, and 68% to total WSIs, respectively. Among individual SIA components, nss-SO₄²⁻ exhibited the highest concentration, averaging 2.9 µg m⁻³ (42% of total WSIs), followed by NH₄⁺ (1.1 µg m⁻³, 16%) and NO₃⁻ (1.0 µg m⁻³, 15%). The seasonal pattern of nss-SO₄²⁻ mirrored that of total SIA, with maximum levels in summer (4.1 µg m⁻³) and minimum levels in autumn (1.9 µg m⁻³). Interestingly, winter concentrations were also elevated (2.5 µg m⁻³), likely due to increased SO₂ emissions and the dominance of heterogeneous aqueous-phase oxidation pathways under cold and humid conditions [76,77].

Marine-related ions, such as Na⁺ and Mg²⁺, were primarily sourced from sea spray and, to a lesser extent, from crustal materials. In our study, sea-salt-derived Mg²⁺ (ss-Mg²⁺) contributed up to 85% of total Mg²⁺, with peak concentrations observed in summer (particularly in July). The Cl⁻-to-Na⁺ ratio ranged from 0.2 to 1.8 across seasons, with the highest values observed in winter (1.6 in PM₁₀ and 1.8 in PM_2.5), followed by spring (1.2 in PM₁₀ and 0.9 in PM_2.5), autumn (0.9 in PM₁₀ and 0.6 in PM_2.5), and summer (0.7 in PM₁₀ and 0.2 in PM_2.5). The lowest ratios, especially in summer, fell well below the seawater reference value of 1.8, indicating chloride depletion via acid displacement reactions (e.g., NaCl + HNO₃ → NaNO₃ + HCl↑) due to interactions with atmospheric nitric and sulfuric acids [78].

Potassium (K⁺), a typical tracer of biomass burning, exhibited higher concentrations in winter, supporting increased biomass combustion during that period [79]. This trend is corroborated by elevated OC-to-EC ratios in winter (2.3 in PM₁₀ and 2.7 in PM_2.5), as presented in Table 2. Moreover, Calcium (Ca²⁺), commonly associated with crustal dust [80], also showed distinct seasonal variations between the two size fractions. The highest Ca²⁺ concentrations were recorded in winter for PM₁₀ and in summer for PM_2.5. Non-sea-salt calcium (nss-Ca²⁺) accounted for over 95% of total Ca²⁺ in both PM fractions, confirming its predominantly terrestrial origin. Despite the evident Mediterranean influence on the aerosol profile [15], the dominance of nss-Ca²⁺ and wintertime K⁺ spikes, combined with high OC-to-EC ratios, underscore the significant role of local sources, particularly dust resuspension and biomass burning.

3.3. Levels and Temporal Distribution of Carbonaceous Compounds in PM₁₀ and PM_2.5

3.3.1. Organic Carbon and Elemental Carbon

As shown in Figure 4 and Figure S4, organic carbon (OC) and elemental carbon (EC) concentrations exhibited distinct seasonal and size-fraction-dependent patterns. In PM_2.5, OC concentrations ranged from 0.10 to 18.2 μg m⁻³, with an annual average of 5.4 ± 2.8 μg m⁻³, while EC levels varied from 0.5 to 8.4 μg m⁻³, averaging at 3.2 ± 1.6 μg m⁻³. Monthly OC concentrations (Figure S5) were highest between November and February (monthly averages: 6.5–8.9 μg m⁻³), peaking in January at 9.6 μg m⁻³. In contrast, EC concentrations showed episodic peaks in July (4.4 μg m⁻³) and September (4.1 μg m⁻³), but remained slightly higher on average during the cold season (~1.1-fold) due to enhanced accumulation under stagnant conditions. In PM₁₀, average OC and EC concentrations were 5.4 ± 2.8 μg m⁻³ and 3.0 ± 1.2 μg m⁻³, respectively, with peak OC observed in January (9.2 μg m⁻³) and minimum values recorded in August (3.9 μg m⁻³ for OC) and May (1.9 μg m⁻³ for EC). These low concentrations represent approximately 2.3-fold differences from the seasonal peaks (Figure S6). Seasonal variations in OC and EC concentrations were statistically evaluated using an ANOVA statistical test. Seasonal ANOVA confirmed significant OC variability across seasons and between cold (autumn and winter) and warm (spring and summer) periods (p < 0.001) for both PM_2.5 and PM₁₀, while EC showed notable seasonal and monthly variation only in PM₁₀ (p < 0.001). During the cold period, OC concentrations increased by approximately 1.8 times in PM_2.5 and 1.5 times in PM₁₀ compared to the warm period. These patterns suggest that the elevated OC levels are likely associated with additional seasonal sources or enhanced formation processes, beyond traffic-related emissions, which tend to be more stable year-round [15,23].

Correlation analysis (Figures S7 and S8) revealed a weak relationship between EC and OC, with correlation coefficients of r = 0.10 in PM_2.5 and r = 0.27 in PM₁₀. This suggests different dominant sources or atmospheric processes for these two carbonaceous fractions. However, a strong correlation was found between OC and secondary organic carbon (SOC) in both size fractions (r > 0.9, p-value < 0.001), indicating that the high contribution of SOC to total OC may be responsible for the weak EC-OC correlation [81]. Similarly, EC was strongly correlated with primary organic carbon (POC), highlighting the influence of primary emissions, such as traffic, on EC levels in the study area.

Meteorological conditions appear to be a key driver of the seasonal variability in OC. OC correlated negatively with ambient temperature (r = −0.56 in PM_2.5, p < 0.001; r = −0.39 in PM₁₀, p < 0.005), supporting the role of low temperatures in enhancing gas-to-particle partitioning and secondary organic aerosol (SOA) formation [23,30]. While no statistically significant correlation was observed between OC and boundary layer height (BLH). The physical effect of shallow winter BLH (280–846 m) likely promoted accumulation and SOC formation under reduced dispersion [30]. In contrast, EC concentrations showed no significant correlation with meteorological variables, such as wind speed, rainfall, relative humidity, or boundary layer height across both PM size fractions, suggesting a more consistent and primarily local origin. Nevertheless, long-range atmospheric transport appears to significantly impact the temporal variability of carbonaceous aerosols at the study site (Figure 5 and Figure S9). During the warm season, regional and transboundary sources contributed to elevated EC episodes despite stable local emissions. During the cold season, when local sources are expected to dominate, yet episodic transport of polluted air masses likely contributed to the high OC and EC levels observed. This combined influence of local meteorological conditions, secondary formation, and the strength of regional transport aerosols may explain the pronounced seasonality of carbonaceous aerosols in Tetouan [82].

3.3.2. Secondary Organic Carbon

The estimated concentrations of primary organic carbon (POC) and secondary organic carbon (SOC), as well as their contributions to total organic carbon (OC), are presented in Table 2, based on the methodology previously described. In PM_2.5, the average concentrations of POC and SOC were 0.4 ± 0.2 μg m⁻³ and 5.0 ± 2.8 μg m⁻³, respectively, corresponding to 2% and 30% of the total PM_2.5 mass. In PM₁₀, SOC levels averaged 4.3 ± 2.2 μg m⁻³ (accounting for 15% of PM₁₀ mass), ranging from 0.8 to 12.5 μg m⁻³, while POC concentrations varied between 0.6 and 3.6 μg m⁻³, with a mean of 1.5 ± 0.6 μg m⁻³ (5% of PM₁₀ mass). Notably, both the mean SOC concentrations and the SOC-to-OC ratios were consistently higher in PM_2.5 than in PM₁₀, indicating a stronger contribution of secondary carbonaceous species in the fine aerosol fraction. SOC accounted for 34–99% of the total OC in PM_2.5 (6–79% of PM_2.5 mass) and 25–94% of OC in PM₁₀ (3–51% of PM₁₀ mass). These results align with previous findings by Liu et al. [83], who reported decreasing SOC contributions with increasing particle size. In Southern Italy, Cesari et al. reported that SOC was mainly segregated in PM_2.5 and represented 53–75% of the total OC [84]. Compared to other North African urban environments, the SOC levels observed in this study were relatively elevated. For instance, SOC in PM_2.5 was significantly higher than values reported for Bou-Ismail, Algeria (1.88 μg m⁻³) [16], and in PM₁₀ from Fez, Morocco (2.2 ± 1.2 μg m⁻³) [85].

Seasonal analysis revealed that SOC concentrations were notably higher during the cold season. In PM_2.5, cold-season SOC averaged 6.6 ± 3.0 μg m⁻³ (35% of PM_2.5 mass), compared to 3.5 ± 1.3 μg m⁻³ (21%) in the warm season. In PM₁₀, SOC averaged 5.4 ± 1.9 μg m⁻³ (18% of PM₁₀ mass) during the cold season and 3.4 ± 1.9 μg m⁻³ (11%) in the warm season. Monthly trends (Figures S5 and S6) confirmed that SOC concentrations rise substantially during the cold period, coinciding with higher OC levels. Interestingly, the SOC-to-OC ratios were higher during the cold period than in the warm period, which contrasts with the conventional understanding that SOC formation is enhanced under warmer, more photochemically active conditions and production of secondary organic aerosols [86,87]. This unexpected trend may be explained by the shallow boundary layers (280–846 m) and a lower ventilation coefficient (the product of WS and BLH) during the cold season (1890 m² s⁻¹), limited vertical dispersion compared to the warm season (2375 m² s⁻¹). Combined with lower temperatures, these conditions likely favoured the accumulation of primary pollutants and enhanced condensation or adsorption of precursors onto pre-existing particles [40]. Galindo et al. [88] suggested that the prevailing meteorological conditions during the cold season could govern the seasonal SOC cycle. While the SOC-to-OC ratio showed relatively stable daily values within each month (p-value > 0.05, one-way ANOVA), significant monthly and seasonal variations were observed (p-value < 0.01). Table 2 summarizes seasonal averages of SOC-to-OC ratios. In PM_2.5, winter recorded the highest average ratio (0.94), followed closely by spring (0.92), autumn (0.93), and summer (0.85). A similar seasonal pattern was observed in PM₁₀, with a maximum ratio of 0.77 in winter and a minimum of 0.64 in summer. Similar results were observed in Lecce, Italy, where SOC-to-OC was found to be 0.8 in winter compared to 0.6 in summer [84].

Overall, SOC contributions exhibited clear seasonal and size-dependent patterns, with winter dominance in both fractions. SOC contributions in Tetouan are among the highest reported for Mediterranean urban areas. Wintertime SOC in PM_2.5 reached 35% of total PM_2.5 mass, exceeding levels observed in Athens, Greece (7.5%; [89]) and Lecce, Italy (25.8%; [84]), likely due to enhanced partitioning of semi-volatile organics at low temperatures (≤13 °C, Table 2), shallower boundary layers (BLH: 280–846 m), and reduced ventilation (1890 m² s⁻¹). Although photochemical production is typically higher in summer, wintertime conditions favour the condensation and adsorption of semi-volatile organics, leading to enhanced SOC accumulation. This seasonal behaviour aligns with findings from other urban sites, such as Beijing [40], Mongolia [90], and Hong Kong [91], further reinforcing the interpretation that local accumulation processes dominate in winter, while summer SOC is more influenced by local photochemical production and episodic transport from regional sources [41].

3.3.3. Total Carbonaceous Aerosol

The analysis of total carbonaceous aerosol (TCA) concentrations at the Tetouan site revealed annual averages of 9.9 ± 3.4 μg m⁻³ for PM₁₀ and 9.6 ± 3.7 μg m⁻³ for PM_2.5. TCA constituted a substantial fraction of the total particulate mass, accounting for approximately 52% of PM_2.5 and 34% of PM₁₀, highlighting a notable enrichment of carbonaceous species in the fine particle fraction. Temporal patterns of TCA closely followed those observed for organic carbon and secondary organic carbon, as depicted in Table 2 and Figure 4, Figures S5, and S6. However, the relative contribution of TCA to total PM mass exhibited distinct seasonal variations. In PM_2.5, the TCA contribution remained consistently elevated across all seasons, ranging from a maximum of 60% in autumn to a minimum of 44% in spring. In contrast, the relative contribution of TCA-to-PM₁₀ exhibited a clear seasonal cycle, peaking in autumn (45%) and reaching a minimum in summer (28%). These seasonal differences may be attributed to both source activity and meteorological conditions. Specifically, elevated TCA levels are likely driven by enhanced emissions of carbonaceous species during colder months, combined with unfavourable dispersion conditions, such as low wind speeds, limited atmospheric mixing, and frequent temperature inversions, which facilitate the accumulation of particulate matter [41]. The particularly high autumn contributions likely reflect the onset of carbonaceous-related emissions coupled with increasingly stable atmospheric conditions, which continue to favour accumulation throughout winter.

The TCA contribution to PM_2.5 at the Tetouan site is consistent with observations from other urban and suburban locations. For instance, similar contributions were reported at a residential-commercial site in Belluno, Italy (47%, average TCA-to-PM_2.5) [92], and in urban Chiang Mai, Thailand (50%) [93]. In contrast, significantly lower TCA-to-PM_2.5 ratios have been documented at elevated sites, such as Mount Heng in South China, where TCA accounted for only 20.7% of PM_2.5 mass [94]. For PM₁₀, the carbonaceous contribution observed in Tetouan is comparable to values reported at rural background sites, including Ispra, Italy (36%) and Illmitz, Austria (34%) [95]. Conversely, lower contributions have been found in several European urban areas, such as Ghent, Belgium (25%) and San Pietro Capofium, Italy (28%) [95]. Further research by Herrera Murillo et al. [96] within the Costa Rican Metropolitan area reported average TCA-to-PM₁₀ ratios of 35%, with a range of 28–45% across different commercial, industrial, and residential environments.

3.4. Levels and Variability of Sugar Compound Concentrations

This section provides insight into the temporal distribution of sugar compounds within PM_2.5 and PM₁₀ fractions, specifically within the warm season, covering both spring and summer. Sugar-based compounds in the atmospheric aerosols serve as well-established tracers for biological origin, including primary bio-aerosols and biomass burning emissions [97,98]. This study investigates three representative sugar markers: levoglucosan (an anhydrosugar linked to biomass burning), arabitol (a sugar alcohol emitted by fungal spores), and glucose (a primary saccharide from plant and microbial sources). Table 2 summarizes the seasonal averages and concentration ranges of these species, while Figure 6 shows their seasonal trends in PM_2.5 and PM₁₀, respectively. Monthly variations are illustrated in Figures S10 and S11.

3.4.1. Levoglucosan

Levoglucosan is often used as a well-established tracer for biomass-burning activities [38,99] due to its remarkable atmospheric stability and its distinct emission pathway [100]. In this study, levoglucosan exhibited notable differences between PM_2.5 and PM₁₀ fractions. In PM_2.5, levoglucosan concentrations ranged from 0.5 to 49.2 ng m⁻³, with an average of 11 ± 12 ng m⁻³, while in PM₁₀, levels were higher, averaging 22 ± 17 ng m⁻³ (range: 1.5–84.0 ng m⁻³). A clear seasonal pattern was evident, with higher concentrations recorded during spring (16.6 ng m⁻³ in PM_2.5 and 26.7 ng m⁻³ in PM₁₀), followed by a marked decline in summer months (5.5 ng m⁻³ in PM_2.5 and 17.2 ng m⁻³ in PM₁₀). This seasonal behaviour aligns with observations at the Finokalia station in the Eastern Mediterranean [101]. It is likely driven by enhanced photochemical degradation of levoglucosan by hydroxyl (OH) radicals during summer, when solar radiation and oxidant levels are elevated [100]. Moreover, the temporal trend of OC showed a significant correlation with levoglucosan (r = 0.5 in PM₁₀, p < 0.001), confirming the substantial influence of biomass combustion on carbonaceous aerosol loading [102]. The pronounced spring peak likely reflects a combination of regional burning and local biomass use, occurring before enhanced summer photochemical removal processes become dominant. Furthermore, as illustrated in Figures S7 and S8, a moderate but statistically significant positive correlation was observed between levoglucosan and K⁺ in both PM_2.5 (r = 0.5, p < 0.05) and PM₁₀ (r = 0.5, p < 0.001). This consistency supports their common origin from biomass burning, in agreement with previous findings linking these two markers to the combustion of biomass [38,99]. However, the moderate correlation strength (r = 0.5) suggests that additional factors may influence their relative abundances.

Levoglucosan concentrations measured at the Tetouan urban site exhibited a two-to-five-fold decrease compared to documented values in other Mediterranean regions. This difference can be attributed to a combination of local emission patterns and atmospheric degradation processes. Unlike many European urban centres where residential wood burning for heating is a dominant wintertime source, the scale and type of biomass combustion in Tetouan may be different [103]. Furthermore, levoglucosan is known to undergo photochemical and oxidative degradation in warm and sunlit environments, which can significantly decrease its atmospheric lifetime [104,105]. This process, combined with potential differences in local source strength, likely contributes to the lower ambient concentrations when compared to other cities where colder temperatures and a higher prevalence of biomass heating create conditions for greater accumulation and longer atmospheric persistence of levoglucosan. For instance, PM₁₀ levoglucosan concentrations in Tetouan (21.7 ± 16.9 ng m⁻³) were significantly lower compared to reported values in a coastal Algerian site (Bou Ismail, 52.3 ng m⁻³) [106], an urban Spanish city (Barcelona, 60 ng m⁻³) [107], a rural Andalusian location in springtime (100 ng m⁻³) [108], and in an urban area in Portugal (Oporto, ~120 ng m⁻³ in winter) [109]. Similarly, in PM_2.5, levoglucosan concentrations in Tetouan (11 ± 12 ng m⁻³) were well below those reported for Beirut, Lebanon (49 ng m⁻³) [110] and Aveiro, Portugal (58 ng m⁻³) [99]. The levels observed in Tetouan were, however, comparable to recent findings from Fez, Morocco (27 ± 15 ng m⁻³ for PM₁₀) [85]. In contrast, this same study also documented exceptionally low concentrations (2.0 ± 1.1 ng m⁻³) for a remote mountainous site in the Middle Atlas, further emphasizing the importance of local emission sources such as biomass combustion in determining urban levoglucosan levels [85].

While previous studies show that levoglucosan is predominantly associated with fine particles (less than 2 µm) in urban environments [99,111,112], our findings indicate a substantial shift, with the majority of levoglucosan mass observed in the coarse aerosol fraction (PM₁₀). This hitherto unreported shift in the size distribution of levoglucosan within urban settings can potentially be attributed to two key mechanisms. First, the adsorption of semi-volatile organic compounds, including levoglucosan, onto pre-existing mineral dust particles could facilitate their presence in the coarse mode [109]. Second, the pronounced polarity and hygroscopicity of biomass burning-derived organics may promote aqueous-phase processing and condensational growth, ultimately shifting the particle size distribution toward larger diameters under humid atmospheric conditions [109,113].

3.4.2. Arabitol

Among sugar alcohols, arabitol is widely recognized as a molecular marker for primary biogenic emissions, particularly from fungal spores [114,115]. Its presence in the atmosphere is mainly attributed to the resuspension of dust or soil particles enriched with biological materials, including fungal hyphae, spores, and bacterial colonies [116]. These emissions are often linked to natural soil erosion, agricultural activity, and traffic on unpaved roads, all of which disturb biota-rich surfaces and contribute to airborne particulate matter [117]. In the present study, arabitol concentrations were consistently higher in the PM₁₀ compared to the PM_2.5, reflecting the typical aerodynamic size of fungal spores. PM_2.5 samples exhibited relatively low concentrations, with an average of 0.23 ± 0.26 ng m⁻³ and a range from below detection to 1.26 ng m⁻³. Conversely, significantly higher concentrations were observed in PM₁₀, averaging 1.9 ± 2.6 ng m⁻³ and ranging up to 11.7 ng m⁻³. This strong coarse-mode dominance aligns with the size distribution of primary bio-aerosols and their frequent association with mineral dust particles [117].

A seasonal spring peak in arabitol was observed in PM₁₀, while PM_2.5 concentrations remained relatively stable between spring and summer (Figure 6, Figures S10 and S11). This pattern suggests that the onset of biological activity in spring, combined with enhanced dust resuspension and favourable meteorological conditions, contributes to higher coarse-mode arabitol levels. Although summer temperatures are higher and continue to support biological emissions, the lack of a further increase likely reflects reduced dust resuspension under drier, less disturbed surface conditions. These observations are consistent with previous findings that temperature, thermal convection, and light-induced photophoresis (light-induced movement) can facilitate fungal spore release [118]. Petit et al. further attributed such seasonal trends to the temperature-dependent emission of fungal spores into the atmosphere, which are known precursors of sugar alcohols such as arabitol [114]. Regionally, Tetouan arabitol concentrations are significantly lower than those reported in other locations. For instance, Deabji et al. measured mean springtime concentrations of 9.9 ng m⁻³ in urban Fez, Morocco [85], while Ledoux et al. reported 10.7 ng m⁻³ during winter at a rural coastal site in Northern France [119].

3.4.3. Glucose

Glucose, a sugar compound primarily originating from plant detritus, such as pollen, fruit, and leaves [100,101,120], exhibited a wide range of daily concentrations in this study. In PM₁₀, glucose concentrations ranged from 0.1 to 67.7 ng m⁻³, with an average of 10 ± 15 ng m⁻³, while in PM_2.5, they varied from 0.1 to 9.6 ng m⁻³, with an average of 2.2 ± 2.4 ng m⁻³. These values are slightly lower than those reported in previous studies, such as at a rural Greek site (0.48–110 ng m⁻³; mean: 13.5 ng m⁻³ in PM₁₀) [101] and at European urban traffic sites (13.6 ± 12.6 ng m⁻³ in PM₁₀) [120]. These results suggest that glucose is predominantly associated with coarse particles, consistent with prior findings indicating its abundance in the PM₁₀ fraction [100,120]. However, Carvalho et al. [121] reported spatially dependent shifts in glucose size distribution, with either fine- or coarse-mode dominance depending on sampling site characteristics.

Further investigation into the potential factors influencing glucose concentrations revealed divergent behaviours between the PM_2.5 and PM₁₀ fractions. In the PM_2.5, no significant correlations were identified between glucose levels and daily meteorological variables. In contrast, PM₁₀ glucose concentrations exhibited a more complex relationship with meteorological conditions. Multiple linear regression (MLR) analysis was carried out and demonstrated a strong positive association between glucose concentrations and rainfall, alongside negative associations with relative humidity (RH) and boundary layer height (BLH). This relationship is described by the regression equation (Equation (6)) with a robust fit (r = 0.92). Indeed, the MLR analysis underlines the dominant role of meteorology in coarse-mode glucose dynamics.

Glucose (PM₁₀) = 2.7 × Rainfall − 0.14 × RH − 0.12 × BLH + 74.0

(6)

On the other hand, temporal variability analysis revealed distinct patterns between the two particle size fractions. Within PM₁₀, glucose concentrations peaked during early spring, reaching a seasonal average of 14.1 ng m⁻³. However, the PM_2.5 fraction exhibited a more balanced seasonal distribution, with average concentrations of approximately 2.0 ng m⁻³ across both spring and summer. Despite this apparent seasonal stability, a closer examination of monthly data (Figures S10 and S11) revealed a significant inter-monthly variation. Remarkably, July recorded the highest glucose concentration in PM_2.5 (4.7 ng m⁻³), substantially exceeding levels in other months, which ranged from 0.48 ng m⁻³ in September to 2.7 ng m⁻³ in May. This pronounced inter-monthly variability suggests that factors beyond seasonal meteorological patterns may be at play. These could include variations in biological emissions, local anthropogenic activities, or changes in air mass origin (as indicated in Figure 7, upper panel). Figure 7 presents the Concentration Weighted Trajectory (CWT) analysis of glucose levels, highlighting their variability as a function of air mass origin, during May and July for PM_2.5, and during summer and spring for PM₁₀ (lower panel). The CWT analysis further reinforces the idea that long-range transport from the Mediterranean basin and the Iberian Peninsula seems to play a role in the inter-monthly and inter-seasonal variations of glucose in both PM_2.5 and PM₁₀. However, it is essential to consider the potential contributions of local sources. These findings are consistent with those of Theodosi et al. [101], who reported peak concentrations of glucose and other primary sugar compounds during spring, followed by a decline as the growing season progressed. They also emphasized that the concentrations of these compounds can be significantly influenced by the origin of air masses.

4. Conclusions

Continuous sampling of the PM_2.5 and PM₁₀ aerosol fraction on a daily basis was conducted at an urban site in Tetouan (Northern Morocco, Southwestern Mediterranean) over a one-year period (2011–2012). Subsequent chemical analyses allowed the determination of organic carbon (OC), elemental carbon (EC), water-soluble inorganic ions (WSIs), and sugar compounds concentrations, and the estimation of secondary organic carbon (SOC). This investigation aimed to gain a better understanding of the influence of the season and air mass origin on the chemical composition of both aerosol fractions. The main conclusions arising from this study are summarized below:

• PM_2.5 and PM₁₀ in Tetouan exhibit pronounced seasonal variability, with higher concentrations during winter and autumn and lower levels in spring. Annual mean concentrations were 18.0 ± 6.5 µg m⁻³ for PM_2.5 and 30.8 ± 9.7 µg m⁻³ for PM₁₀, with an average PM_2.5-to-PM₁₀ ratio of 0.59, highlighting the dominance of fine particles, particularly during colder periods.
• WSIs contributed ~39% of PM_2.5 and ~44% of PM₁₀ mass, with maximum levels in summer, likely driven by enhanced photochemical activity. Non-sea-salt sulfate showed the strongest summer maximum (4.1 µg m⁻³ in PM_2.5), while nitrate and chloride exhibited winter peaks (1.5 µg m⁻³ and 0.7 µg m⁻³ in PM_2.5; 3.0 µg m⁻³ and 1.6 µg m⁻³ in PM₁₀, respectively), reflecting seasonal shifts in gas-to-particle partitioning under varying temperature and humidity conditions. Seasonal variations indicate that cold periods favour the accumulation of locally emitted and transported coarse-mode salts, while warm periods enhance the formation of secondary inorganic aerosols through atmospheric chemical processes. Therefore, a dual influence of local sources (winter) and regional photochemistry (summer) in controlling the seasonal dynamics of WSIs in the Southwestern Mediterranean.
• Carbonaceous components were the main contributors to fine aerosol mass, with total carbonaceous aerosol (TCA) accounting for ~52% of PM_2.5 and ~34% of PM₁₀. OC concentrations were significantly higher during colder months, driven by both meteorological conditions and enhanced secondary organic carbon (SOC) formation. EC showed a less pronounced seasonal trend, reflecting its primary origin from combustion sources, which tend to have relatively constant emissions throughout the year. SOC was particularly abundant in PM_2.5, reaching 7.3 ± 3.4 µg m⁻³ in winter and contributing up to 90% of OC, reflecting the significance of secondary aerosol formation processes, especially under low-temperature and low-mixing-height conditions. Although local meteorology was a major driver during the cold season, long-range atmospheric transport and photochemistry contributed noticeably to the warmer period. The elevated SOC-to-OC ratios observed in winter were strongly associated with thermodynamic conditions, namely low temperature and limited vertical mixing, whereas in summer, the influence of these meteorological parameters was minimal, and the increase in SOC was attributed to enhanced photochemical oxidation and episodic transport from regional sources.
• Sugar tracers, including levoglucosan, arabitol, and glucose, exhibited pronounced seasonal and size-dependent patterns, with a clear coarse-mode dominance and spring maxima. The unexpected coarse-mode enrichment of levoglucosan (21.7 ± 16.9 ng m⁻³ in PM₁₀) suggests either adsorption onto larger particles or growth through hygroscopic processes. Levoglucosan decline in summer indicates enhanced photochemical degradation by OH radicals under high solar radiation. Arabitol, primarily emitted by fungal spores, was strongly associated with PM₁₀ (1.9 ± 2.6 ng m⁻³) and exhibited spring maxima consistent with periods of intensified biological activity and favourable thermal convection. Glucose, originating from plant detritus and microbial material, also showed coarse-mode dominance, with PM₁₀ concentrations peaking in early spring (14.1 ± 18.9 ng m⁻³) and inter-monthly variability in PM_2.5, with a notable July peak (4.7 ng m⁻³).

Overall, the findings offer new insights into the size-resolved chemical composition and seasonal dynamics of aerosols in a North African urban setting. The results highlight the interplay between local emissions, meteorological conditions, and regional atmospheric transport in shaping the chemical profile of PM₁₀ and PM_2.5. Although the highest CWT values were observed over the Mediterranean Sea, this does not necessarily imply that the sea itself is a dominant source of PM. Instead, these elevated values primarily reflect the convergence of air masses carrying pollutants from continental sources (Europe, North Africa, and the Iberian Peninsula) as they transit over the sea before reaching Tetouan. The Mediterranean acts mainly as a transport corridor, with contributions from marine aerosols and maritime shipping, while the apparent hotspots are largely due to trajectory confluence rather than local emissions. While CWT analysis effectively resolved regional PM sources, future studies could enhance precision by integrating local emission inventories (e.g., traffic, industrial) and vegetation phenology data (e.g., dust/biogenic emissions). Such hybrid approaches would better partition local and transported contributions, particularly in regions with complex source mixtures. This work reinforces the characterization of aerosols in Tetouan as exhibiting a mixed urban–Mediterranean aerosol signature, influenced by both local activities and regional sources.