The Impact of Intense Winter Saharan Dust Events on PM and Optical Properties at Urban Sites in the Southeast of the Iberian Peninsula

Abstract

The influence of three Saharan dust events (SDE) on particulate matter (PM) concentrations and aerosol optical properties (AOP) during February 2021 was studied. The physical characteristics of the African aerosol were different for each episode. Therefore, the impacts of the three events on PM and AOP were analyzed separately. The monitoring sites were placed in Elche, in the southeast of the Iberian Peninsula. The sites can be classified as urban background locations. The procedure used to obtain the contribution of SDE to PM₁₀ mass concentrations was the 40th percentile method. Nearly half of the days during the study period were under the influence of Saharan air masses. The average contribution of mineral dust (MD) to the PM₁₀ mean concentration was ~50%, which was the highest contribution during the month of February in the last 14 years. The results show that those events characterized by a high input of fine particles (PM₁ and PM_2.5) caused larger increases in the absorption (σ_ap) and scattering (σ_sp) coefficients than SDE in which coarse particles predominated. Nevertheless, as expected, SAE (Scattering Angström Exponent) values were lowest during these episodes. AAE (Absorption Angström Exponent) values during SDE were slightly higher than those observed in the absence of African dust, suggesting some contribution from MD to the absorption process.

1. Introduction

The Mediterranean Basin is exposed to continuous incoming air masses loaded with dust from North Africa. The influence of Saharan dust transport on particulate matter (PM) levels has an expected latitudinal dependence [1,2]. Therefore, the impact of these episodes on PM levels in southern regions within the European continent is significantly higher than that observed in more northerly areas.

The south-eastern Iberian Peninsula (SEIP), located in the Western Mediterranean Basin (WMB), is geographically close to North Africa, hence the occurrence of Saharan dust events (SDE) in this region is quite common. Several studies have determined that the annual frequency of these events in the SEIP ranges between 22 and 26% [1,3]. At regional background (RB) stations, an 18–25% contribution of mineral dust (MD) to annual mean PM₁₀ values (i.e., particles with aerodynamic diameters < 10 μm) has been estimated [1,3,4]. Although during the winter period the frequency of these events is lower (~14%), the MD contribution to PM₁₀ is similar to the annual average [3]. It has been established that the mean annual number of exceedances of the daily PM₁₀ limit value (2008/50/CE Directive) in this area is ~4 [5].

In urban areas located in the SEIP, the impact of SDE is greater on coarse particles than on fine particles, not only increasing their concentration, but also modifying their chemical composition [6,7]. The PM₁₀ fraction can be considered as the best air quality indicator for evaluating the short-term health effects of Saharan dust [8]. This is supported by the results of previous works. For instance, Pérez et al. [9] observed an increase in daily mortality in the city of Barcelona during SDE due to the increase of coarse PM concentrations. On the other hand, in a study carried out in Madrid [10], the authors concluded that an increase of 10 μg·m⁻³ in the coarse fraction (PM_10–2.5) during Saharan dust days was associated with a 2.8% increase in total mortality, compared with 0.6% during non-dust days.

In addition to its adverse effects on human health, MD can provide a surface site for heterogeneous reactions of atmospheric gases, and can also absorb and scatter solar radiation. Therefore, MD may lead to a change in the radiative balance. It is generally established that MD has a slight net cooling effect, although with a high level of uncertainty. Its direct radiative forcing varies between −0.3 and +0.1 W·m^–2 at a global scale [11]. Aerosol optical properties (AOP) can undergo significant variations under the influence of SDE. Several studies carried out in the Iberian Peninsula have analyzed the impact of SDE on AOP [12,13,14,15,16,17]. The main variations in AOP during SDE include: increases in the scattering (σ_sp) and absorption (σ_ap) coefficients; a rise in the Absorption Angström Exponent (AAE) due to the absorption of UV and visible radiation by MD; a drop in the Scattering Angström Exponent (SAE) because of the predominance of coarse particles during these events; and a rise in the Single Scattering Albedo (SSA) value since the relative increase in σ_sp is greater than that in σ_ap. On the other hand, AOP variations depend on the physicochemical properties of MD. MD is mainly made up of quartz and silicate minerals. Other main components (whose contribution depend on the source region) include organic matter, Fe₂O₃, Al₂O_3, and CaCO₃. MD is typically irregular in shape and is characterized by a large size spectrum, from hundreds of nanometers to tens of micrometers [18]. MD can undergo different aging processes. Therefore, its chemical and physical properties may vary during transport, leading to changes in its optical properties [18]. In addition, the radiative effect of MD may be affected by other factors such as the travel distance, the pathway from source regions, and the atmospheric lifetimes of the particles [19].

Size distribution is an essential characteristic for estimating the aerosol radiative effect, but its representation remains challenging due to the large size spectrum of MD [20]. In order to gain insights into the influence of aerosol size distributions on AOP, this work is focused on the analysis of the variations in AOP during three different SDE. Additionally, the contribution of SDE to PM₁₀ levels in an urban area located in the SEIP during wintertime was quantified.

2. Materials and Methods

2.1. Sampling Sites

Two urban background (UB) sites located in Elche (38°16′ N; 0°41′ W; 86 m.a.s.l) were selected for this work. The city is located in the south-eastern Spanish Mediterranean basin, about 12 km from the coast, and is highly sensitive to SDE due to its proximity to the African continent. Recent work addressing the impact of Saharan mineral dust on PM levels and composition in this area can be found in [6,7,21,22].

Data from the “Agroalimentari (AGRO) (38°14′ N; 0°41′ W; 54 m.a.s.l)” station (see Figure 1), located on the outskirts of the city, were used in the first part of this study (Section 3.1.1). In this section, the MD load at an RB station was calculated in order to quantify the impact of SDE on PM₁₀ levels over the last 14 years. The selected station was located in the town of Pinoso (PIN) (38°25′ N; 1°02′ W; 637 m.a.s.l.), about 44 km from Elche. Both stations belong to the Environmental Surveillance Network of the regional Government of Valencia. A monitoring station equipped with suitable optical instruments and located at the Miguel Hernández University of Elche (UMH) (38°16′ N; 0°41′ W; 93 m.a.s.l) was chosen to study the variation in AOP under SDE during February 2021 (Section 3.1.2 and Section 3.2). At this station, PM₁₀, PM_2.5, and PM₁ concentrations were continuously measured using a GRIMM 190 optical counter. The spectrometer provided PM levels with a 10 min time resolution. One hour and twenty-four hour averages were subsequently calculated. Daily values were corrected by comparison with PM concentrations obtained by the gravimetric technique.

Table 1. Measurements of PM levels and/or AOP at the three monitoring stations.

Station (Type)	Measurements	Time Resolution	Time Period	Purpouse b
AGRO (UB)	PM10	Daily	Feb. (2008–2021)	SDE on PM10
PIN (RB)	PM10	Daily	Feb. (2008–2021)	SDE on PM10
UMH (UB)	PMx a, AOP	Daily/Hourly	Feb. (2021)	SDE on AOP

^a: PM₁₀, PM_2.5 and PM₁; ^b: To study the influence of SDE on PM₁₀ or AOP.

summarizes the measurements performed at each site.

2.2. SDE Detection

The identification of SDE was based on the results of the predictive model BSC-DREAM8b (https://ess.bsc.es/bsc-dust-daily-forecast, accessed on 3 November 2021). The analysis of PM₁₀ time series helped to confirm the occurrence of SDE over the study area. Meteorological scenarios were investigated based on synoptic charts at 850 hPa geopotential height from NCEP reanalysis with a spatial resolution of 2.5° × 2.5°.

The procedure used to estimate the MD load in PM₁₀ during SDE was the “P40 method” http://ec.europe.eu/environment/air/quality/legislationpdf/sec_2011_0208.pdf (accessed on 3 November 2021). The basis of this method can be found in [23,24]. A good summary of this methodology can also be found in [25]. Briefly, the method consists of calculating the monthly moving 40th percentile of the PM₁₀ time series at an RB site after excluding days with desert dust outbreaks. This is then denoted as the RB level. The daily net contribution of MD to PM₁₀ levels in a given region can subsequently be obtained by subtracting the daily RB level from the daily PM₁₀ concentration measured at the RB site [23]. The P40 method is commonly used to determine the MD contribution to PM₁₀ and PM_2.5 where chemical speciation is not required. In addition, it should be noted that, although the calculation was done for the RB stations, this methodology allows estimating the daily contribution of African dust for the whole region [25].

2.3. Aerosol Optical Properties

Scattering (σ_sp) and backscattering (σ_bsp) coefficients at 450, 525, and 635 nm were obtained by means of an LED-based integrating nephelometer (model Aurora 3000, ECOTECH Pty Ltd., Knoxfield, Australia). A full calibration of the nephelometer was performed four times per year by using CO₂ as span gas, while zero check measurements were performed daily by using internally filtered particle free air. Scattering coefficients were corrected for truncation and for non-Lambertian errors according to the method reported by Müller et al. [26]. In this work, correction factors for total scattering at each wavelength were obtained from Angström exponents using the parameters provided by Müller et al. [26]. The measurements were conducted under dry conditions using a processor-controlled automatic heater. The time resolution was 5 min. Hourly averages were subsequently calculated. σ_sp and σ_bsp were measured during February 2021.

The backscatter fraction (b) was obtained from the following equation:

b = σ_bsp/σ_sp

(1)

The scattering Angström exponent (SAE) was derived from σ_sp values as follows:

$$\mathrm{SAE}=-\frac{\ln \frac{{\mathsf{\sigma }}_{\mathrm{sp}}{(\mathsf{\lambda }}_1)}{{\mathsf{\sigma }}_{\mathrm{sp}}{(\mathsf{\lambda }}_2)}}{\ln \frac{{\mathsf{\lambda }}_1}{{\mathsf{\lambda }}_2}}$$

(2)

where λ₁ = 450 nm and λ₂ = 635 nm.

The scattering Angström exponent describes the wavelength-dependence of scattered light. In situations where the scattering is dominated by fine particles (d < 1 μm), SAE values are usually equal to or greater than 1.5 [27], while values close to 0 occur when the scattering is dominated by coarse particles [28]. Negative values also occur when MD is the dominant aerosol component [29].

Aerosol absorption coefficients (σ_ap) at seven wavelengths (370, 470, 520, 590, 660, 880, and 950 nm) and BC (Black carbon) concentrations were obtained by means of a dual-spot aethalometer (AE33, Magee Scientific, Berkeley, CA, USA). The AE33 minimizes the loading effect [30]. The M8060 filter with a multiple scattering parameter C = 1.39 has been used according to the manufacturer’s instructions. The time resolution was 5 min. Hourly averages were subsequently calculated. σ_ap was measured during February 2021.

The absorption Angström exponent (AAE) was obtained from σ_ap values using the following equation:

$$\mathrm{AAE}=-\frac{\ln \frac{{\mathsf{\sigma }}_{\mathrm{ap}}{(\mathsf{\lambda }}_1)}{{\mathsf{\sigma }}_{\mathrm{ap}}{(\mathsf{\lambda }}_2)}}{\ln \frac{{\mathsf{\lambda }}_1}{{\mathsf{\lambda }}_2}}$$

(3)

where λ₁ = 370 nm and λ₂ = 950 nm.

The AAE value can provide an indication of the aerosol composition. When BC is the dominant absorbing aerosol component, the AAE value is close to 1 [31], while MD aerosols typically have AAE values significantly higher than 1 [29,32].

Finally, SSA values were calculated from σ_ap and σ_sp coefficients:

$$\mathrm{SSA}=\frac{{\mathsf{\sigma }}_{\mathrm{sp}}}{{\mathsf{\sigma }}_{\mathrm{sp}}+{\mathsf{\sigma }}_{\mathrm{ap}}}$$

(4)

The SSA provides information about the cooling or warming effect of atmospheric aerosols.

3. Results and Discussion

3.1. Impact of Dust Aerosols on PM Concentrations

3.1.1. Influence of SDE on PM₁₀ Levels in Wintertime

The presence of Saharan dust over the SEIP is quite constant throughout the year, although the frequency of SDE increases substantially between spring (April, May, June) and autumn (October, November, December) [1]. During the colder months (January, February, March) the frequency and duration of dust events tend to be lower. The average number of days under the influence of Saharan dust events in the SEIP during wintertime is about 13 [3]. However, as previously mentioned, this does not imply that the percentage contribution of Saharan dust to PM₁₀ levels is lower than in other seasons.

Figure 2a shows the relationship between the number of days influenced by Saharan dust intrusions (N_SDE) and the mean PM₁₀ concentration obtained at AGRO during the month of February over the last 14 years (2008–2021). The correlation was statistically significant at a 95% confidence level (R² = 0.75). A background PM₁₀ concentration of 16.5 μg·m⁻³ was obtained on non-event days. As seen in Figure 2a, large variations in the mean PM₁₀ concentrations were registered, with the largest difference (~17 μg·m⁻³) being between 2008 and 2018.

Figure 2a also shows average PM₁₀ levels (blue horizontal line) and N_SDE levels (blue vertical line) obtained during the month of February for the 2008–2014 period. The intersection of both lines divides the graph into two sections. The years in which the N_SDE and the mean PM₁₀ concentration were lower than the average values for the whole study period (22.2 μg/m³ and ~6) fall within the green area of the plot. For those years within the red area of the graph, the N_SDE and the mean PM₁₀ concentration were above the average values for the whole period. The number of exceedances of the PM₁₀ daily limit value (50 μg·m⁻³, 2008/50/EC Directive) due to SDE are shown in brackets. Exceedances caused by other types of events were not considered. For example, during the colder months, episodes of high atmospheric stability leading to the accumulation of pollutants are frequent. Although these events have a greater impact on PM₁ and PM_2.5 concentrations, significant increases in PM₁₀ levels are also observed [6].

Both the determination coefficient and the slope of the line shown in Figure 2a were lower when all winter months were considered (R² = 0.43; slope = 0.28). The reason for this is that using a longer period of data does not result in proportional increases in the N_SDE and the mean PM₁₀ concentrations, possibly because the relative increase in the number of days with air masses coming from the Atlantic (sometimes associated with precipitation events) is greater than the increase in the N_SDE.

To determine the contribution of Saharan dust to PM₁₀ mass concentrations during the month of February between 2008 and 2021, the P40 method was used. The MD load was calculated at the Pinoso station (PIN). Figure 2b shows the MD contribution to mean PM₁₀ levels during February for the whole study period. The RB contribution was determined from PM₁₀ concentrations measured at PIN after subtracting the MD load. On the other hand, the UB contribution was calculated by subtracting the RB and MD contributions to the total PM₁₀ concentration recorded at AGRO.

An increase in the MD load in the second half of the study period was observed, especially during February 2021, when MD contributed ~50% to the average PM₁₀ concentration. On average, the percentage contribution of MD to PM₁₀ levels was ~14%, showing the relevance of this type of event in the SEIP.

3.1.2. Characterization of SDE during February 2021

February 2021 falls within the red area of Figure 2a and had up to 12 days under the influence of Saharan dust (three of them with PM₁₀ levels above 50 μg·m⁻³). Furthermore, as already mentioned, the contribution of MD was the highest of the whole study period. For this reason, SDE occurring during the month of February 2021 were studied in more detail. With this objective, PM and AOP data from the UMH station were analyzed. PM₁₀ concentrations measured at the UMH site showed a high correlation with those recorded at AGRO (PM_10(UMH) = 1.1·PM_10(AGRO); R² = 0.94).

Figure 3 shows the evolution of PM₁₀ daily concentrations as well as the PM₁/PM₁₀ ratio during February 2021. The monthly average PM₁₀ and PM₁/PM₁₀ values were 26.9 μg·m⁻³ and 0.38, respectively. This ratio suggests a large contribution from coarse particulate matter at the study site. The mean PM_2.5 and PM₁ concentrations were, respectively, 15.1 and 9.8 μg·m⁻³. These values are generally low compared with those reported for other urban areas in Europe [33,34]. This is most likely due to the lack of industrial sources in the vicinity of the sampling site and the lower traffic emissions [6].

Figure 3 also shows the periods under the influence of Saharan events (shaded in pink) during February 2021. The third event lasted until the first days of March 2021, although the peak daily concentration was recorded on 27 February. During dust episodes, large increases in PM₁₀ levels were observed, as well as a significant decrease in the PM₁/PM₁₀ ratio. The highest contribution of coarse particles was recorded during the first two episodes, with PM₁/PM₁₀ ratios below 0.2.

Figure S1 (Supplementary Material) shows simulations of dust load (g/m²) and dust opt. depth (550 nm) from the BSC/DREAM8b model over the Iberian Peninsula during the three SDE registered in February 2021. On the other hand, the synoptic patterns (850 hPa) associated with the transport of Saharan dust to the study area were different for the three events, as seen in Figure S1. A low-pressure system located west of the Iberian Peninsula was responsible for the first episode (SDE1). The second episode (SDE2) was characterized by a high-pressure system over the central Mediterranean region coupled with a trough emanating southward from the Icelandic low. Finally, a high-pressure system centerd over the western Mediterranean was the dominant scenario during the third event (SDE3). The characteristics of the three dust episodes registered in February 2021 are presented in

Table 2. Characteristics of the SDE recorded during February 2021 at the UMH site.

	Dates and Duration (N° days)	PM10/PM2.5/PM1 (µg·m−3) a	PM2.5/PM10	PM1/PM2.5	PM1/PM10
SDE1	5–6 (2)	84.8/29.8/12.7	0.33	0.38	0.12
SDE2	17–21 (5)	111.9/49.4/19.9	0.71	0.74	0.55
SDE3	24–28 (5)	44.5/28.4/16.2	0.68	0.65	0.44

^a: Maximum daily values during the three SDE.

.

PM₁₀ levels were highest for SDE1 and SDE2. However, the mass size distribution was quite different during both episodes. A higher contribution of fine particles (<2.5 µm) to PM₁₀ concentrations was observed during SDE2, as shown by the mean values of the PM mass ratios. The contribution of fine particles during SDE3 was also significantly higher than during SDE1.

3.2. Impact of SDE on Optical Properties

3.2.1. General Features

Table 3. Statistical parameters of aerosol optical properties at the UMH station during the study period. b, SAE, SSA, AAE are dimensionless; σ_sp and σ_ap coefficients are given in Mm⁻¹.

	σap (λ = 520 nm)	AAE	σsp (λ = 525 nm)	b	SAE	SSA
Mean	10.1	1.40	53.9	0.16	0.50	0.82
Median	6.3	1.39	35.5	0.15	0.57	0.84
DS	10.2	0.17	55.0	0.04	0.55	0.10
Max	64.9	2.07	355.8	0.53	2.77	0.95
Min	0.3	0.98	1.8	0.10	−1.19	0.36
P5	1.3	1.15	6.9	0.12	−0.46	0.62
P95	31.3	1.69	178.6	0.22	1.22	0.93

presents summary statistics of aerosol optical properties (on an hourly basis) at the UMH site during the study period. The hourly evolution of the measured parameters is shown in Figure S2 (Supplementary Material).

The mean value obtained for σ_ap (10.1 Mm⁻¹) was lower than the values recorded in near-shore urban environments in the Iberian Peninsula such as Burjassot (18.6 Mm⁻¹; λ = 520 nm) [35] and Granada (24 Mm⁻¹; λ = 550 nm) [36]. Similarly, σ_sp values in Burjassot (80 Mm⁻¹; λ = 550 nm) [37] and Granada (61 Mm⁻¹; λ = 550 nm) were higher than that reported here [36]. Nevertheless, comparisons between the aerosol optical properties shown in Table 3 and reference values obtained in previous works at UB sites should be carried out with caution since the monitoring periods are different.

The AAE value (1.40) indicates that, although BC was the main absorbing aerosol component, there was some contribution from non-BC absorbers, such as Brown Carbon (BrC) and MD. Undoubtedly, coarse particles predominated over fine particles and had a prominent role in the scattering process according to the mean SAE value (0.50). In fact, the P95 value (below 1.25) points to the predominance of large particles. This may be due to the following reasons: (1), almost half of the days in the study period were under the influence of African episodes; and (2), coarse particles may be resuspended in the air after the events. In addition, it should be noted that no precipitation events causing wet deposition occurred.

The mean SSA value (0.82) indicates that aerosols were mainly made up of scattering components. However, a certain contribution from absorbing particles cannot be ruled out, as SSA values below 0.85 may indicate the presence of highly absorbing aerosols [38].

3.2.2. MSE and MAE Values

Values of MSE (mass scattering efficiency) and MAE (mass absorption efficiency) were determined from PM mass concentrations and aerosol optical parameters. To obtain MSE and MAE values, the MLR (multilinear regression) method was used. The contribution from submicron (PM₁), fine (PM_2.5–1) and coarse particles (PM_10–2.5) was calculated using the following formulas:

σ_{sp (λ = 520 nm)} = MSE₁·PM₁ + MSE_2.5–1·PM_2.5–1 + MSE_10–2.5·PM_10–2.5 + C

(5)

σ_{ap (λ = 525 nm)} = MAE₁·PM₁ + MAE_2.5–1·PM_2.5–1 + MAE_10–2.5·PM_10–2.5 + C

(6)

C values have no radiative significance [36].

Table 4. Average MSE and MAE (in m²·g⁻¹) for February 2021.

MSE1	MSE2.5–1	MSE10–2.5	R	MAE1	MAE2.5–1	MAE10–2.5	R
4.44 ± 0.06	0.97 ± 0.12	0.35 ± 0.04	0.96	0.61 ± 0.03	−0.18 ± 0.06	0.06 ± 0.02	0.60

presents MSE and MAE values obtained at the UMH site.

σ_ap was mainly contributed by submicron particles, with an insignificant contribution from coarse particles. The values obtained were quite close to those found in the city of Granada [36]. On the other hand, although the greatest contribution to the σ_sp value was from the PM₁ fraction, the influence of larger particles was not negligible. Additionally, it can be observed that the larger the particle size, the smaller the scattering efficiency.

3.2.3. AOP Variations as a Function of the Characteristics of SDE

As shown in Section 3.1.2, particle mass size distributions for the three events detected during the study period were different. To study the impact of each episode on AOP, hourly data were used. Figure 4 shows PM mass concentrations, σ_sp, σ_ap, the extinction coefficient (σ_ext/σ_ext = σ_sp + σ_ap), AAE, SAE, and SSA for each of the three episodes. The values of these parameters in the absence of Saharan events (No SDE) are also shown.

The impact of SDE on PM₁₀ levels was highest during SDE1 (Figure 4a). In contrast, the largest contribution to the fine fractions was observed during SDE2. This affected σ_sp and σ_ap values, as shown in Figure 4b. The median σ_ap value during SDE2 (16.3 Mm⁻¹) was considerably higher than during SDE1 (9.4 Mm⁻¹), and the same was observed for σ_sp. As a result, the highest σ_ext value was recorded during SDE2. SAE values (Figure 4c) were quite low both in the presence and absence of Saharan dust. The average SAE value during SDE1 (0.024) was the lowest due to the large contribution of coarse particles during this event. SSA mean values for the three episodes were very similar and close to 0.880 (Figure 4c). This value was higher than that recorded in the absence of events (0.776). Values close to or above 0.900 (λ = 520 nm) have been previously obtained in the Iberian Peninsula during Saharan events [17,39,40,41]. The spectral variation of SSA can provide useful information about the dominance of specific aerosol types (e.g., BC, organic carbon, dust, sulfates) during events [42,43]. Generally, SSA increases with wavelength for dust aerosols and decreases for biomass-burning and urban/industrial aerosols, while it exhibits almost neutral spectral dependence in the case of the aerosol mixture and sulfate particles [44,45,46]. In addition, an equal mixing of fine and coarse particles is produced when SSA is nearly independent of the wavelength [42]. Figure 5 shows the spectral dependence of SSA (450–635 nm). The slope of the line for SDE1 was significantly higher (almost twice) than for SDE2 and SDE3. This suggests that the contribution of MD was greater during SDE1 than during the other episodes.

Figure 5 also indicates that SDE1 was dominated by coarse aerosols, while during SDE2 and SDE3, the contribution of fine and coarse particles was more similar. An equivalent discrimination of aerosol types according to aerosol size can be found in Patel et al. [42].

The highest mean AAE value was recorded during this first episode (1.525). AAE values during SDE2 and SDE3 were also slightly higher than those obtained in the absence of events. Therefore, it can be assumed that during both SDE, there was some contribution from MD to the absorption process. Figure 6 shows the relationship between AAE values and MD loads on event days (N_SDE = 12). It can be observed that an increment in the MD load was linked with an increase in the AAE value. Similar results were obtained by Ealo et al. [15].

The correlation between AAE and MD concentrations was statistically significant (p-value < 0.05) and indicates that, in the absence of Saharan dust, an AAE value of 1.319 would be obtained. This implies an average MD contribution of 6% to the monthly average AAE value shown in Table 3.

A better characterization of the absorption during these events can be carried out by analyzing the relationship between AAE and EAE (Extinction Angström Exponent). The use of the matrix representing the division of AAE vs. EAE space is widely used [43,46]. This approach provides a good discrimination among light-absorbing PM components (mostly MD, mixed MD and BC, and mostly BC) since each of these species have specific and well-known values of AAE and EAE. In this work, the absorbing aerosols have been distinguished based on the following classification [43]: mostly MD (AAE > 1.8, EAE < 0.8), mostly BC (0.8 < AAE < 1.4, EAE > 1), and mixed MD and BC (AAE ~ 1.5, EAE ~ 0.9).

Table 5. Mean values of EAE and AAE during SDE and in the no SDE period.

	AAE450–635	EAE450–635
SDE1	1.500	0.132
SDE2	1.441	0.766
SDE3	1.429	0.446
No SDE	1.368	0.917

shows the mean values obtained for AAE_450–635 and EAE_450–635 during the three events and for no SDE conditions.

The data shown in Table 5 suggest that during the three SDE, light absorption was produced by mixed MD and BC aerosols. The values for SDE1 were closer to the absorption dominated by MD, while for no SDE conditions, a higher contribution of BC to the absorption process was observed.

4. Conclusions

The average contribution of MD to the PM₁₀ mass concentration at an UB site in south-eastern Spain during February 2021 was ~50%. This value was the highest obtained for the month of February in the last 14 years. The high MD load over the area was due to three SDE detected in February 2021. The characteristics of the three episodes (mass size distributions, duration, synoptic scenarios, etc) were different. SDE1 was characterized by a large contribution from coarse particles, while the highest PM_2.5 and PM₁ mass concentrations were registered during SDE2. As a consequence, the influence of each event on AOP was also different. σ_ap and σ_sp values were highest during SDE2, while the lowest SAE values were obtained for SDE1. The SAE monthly mean value found in this study was quite low (0.50) for an urban environment. A possible reason for this is that coarse particles remained in the atmosphere after dust episodes. MD loads during SDE were significantly correlated with AAE values, indicating a contribution from this component to the absorption process. The relationship between AAE and EAE indicates that the MD contribution was higher during the first event. This outcome is in agreement with that obtained from the SSA spectral analysis, pointing to MD as the dominant aerosol type during SDE1. On the other hand, SSA values were similar during the three events and higher than during non-event periods. This indicates that during SDE, the scattering dominates over the absorption process. From the results obtained it can also be concluded that particles of all sizes are involved in the scattering process, although fine particles scatter light more efficiently (MSE = 4.44 ± 0.06 m²·g⁻¹ for PM₁). In contrast, only the finest particles significantly absorb light.