Analysis and Evaluation of Sulfur Dioxide and Equivalent Black Carbon at a Southern Italian WMO/GAW Station Using the Ozone to Nitrogen Oxides Ratio Methodology as Proximity Indicator

Abstract

The measurement and evaluation of the atmospheric background levels of greenhouse gases (GHGs) and aerosols are useful to determine long-term tendencies and variabilities, and pinpoint peaks attributable to anthropogenic emissions and exceptional natural emissions such as volcanoes. At the Lamezia Terme (code: LMT) World Meteorological Organization–Global Atmosphere Watch (WMO/GAW) observation site located in the south Italian region of Calabria, the “Proximity” methodology based on photochemical processes, i.e., the ratio of tropospheric ozone (O₃) to nitrogen oxides (NO_x) has been used to discriminate the local and remote atmospheric concentrations of GHGs. Local air masses are heavily affected by anthropogenic emissions while remote air masses are more representative of atmospheric background conditions. This study applies, to eight continuous years of measurements (2016–2023), the Proximity methodology to sulfur dioxide (SO₂) for the first time, and also extends it to equivalent black carbon (eBC) to assess whether the methodology can be applied to aerosols. The results indicate that SO₂ follows a peculiar pattern, with LOC (local) and BKG (background) levels being generally lower than their N–SRC (near source) and R–SRC (remote source), thus corroborating previous hypotheses on SO₂ variability at LMT by which the Aeolian Arc of volcanoes and maritime traffic could be responsible for these concentration levels. The anomalous behavior of SO₂ was assessed using the Proximity Progression Factor (PPF) introduced in this study, which provides a value representative of changes from local to background concentrations. This finding, combined with an evaluation of known sources on a regional scale, has been used to provide an estimate on the spatial resolution of proximity categories, which is one of the known limitations of this methodology. Furthermore, the results confirm the potential of using the Proximity methodology for aerosols, as eBC shows a pattern consistent with local sources of emissions, such as wildfires and other forms of biomass burning, being responsible for the observed peaks.

1. Introduction

In the field of atmospheric sciences, various methodologies are used to observe the atmospheric background and its concentration levels of greenhouse gases (GHGs), aerosols, and other parameters. These measurements are necessary to define trends over time, such as the Keeling curve, which is used to determine the increase in carbon dioxide (CO₂) caused by anthropogenic emissions [1,2].

Following the findings of Parrish et al. [3] and Morgan et al. [4], a methodology based on atmospheric photochemical processes, i.e., the ratio of tropospheric ozone (O₃) to nitrogen oxides (NO_x), has been applied at preliminary data gathered by the Lamezia Terme (code: LMT) observation site in Calabria, Southern Italy to assess the source variability of gases such as methane (CH₄) [5]. The results of this early study in LMT observation history allowed for making a number of hypotheses on local sources of pollution, such as livestock farming. Following the findings of another study on the multi-year variability and cyclic patterns of O₃ at the same site [6], a correction factor was introduced to account for summertime peaks in O₃ attributed to regional photochemical pollution [7]. The new correction factor (“ecor”, for enhanced correction) integrated a previous factor (“cor”, correction) based on the overestimation of NO₂ for aged air masses reported by Steinbacher et al. [8] in instruments relying on heated molybdenum converters (~300–400 °C).

Using nine years (2015–2023) of continuous measurements at LMT, the study by D’Amico et al. [7] confirmed the findings of previous research and provided new insights on the source variability of CO (carbon monoxide), CO₂, and CH₄ at the site, thus confirming the potential of the Proximity methodology as an effective tool to discriminate between sources of emissions. However, detailed evaluations based on this method have been limited to the LMT measurements of CO, CO₂, and CH₄ so far, with compounds such as sulfur dioxide (SO₂), which had previously been characterized at the site [9], being excluded from these evaluations. Similarly, no aerosol measured at the LMT station has been evaluated using the Proximity methodology so far, and studies aimed at this type of evaluations are limited [10]. In this study, SO₂ and equivalent black carbon (eBC) will be evaluated for the first time using the Proximity method to test the applicability of such methodology to parameters other than CO, CO₂, and CH₄.

SO₂ can be of anthropogenic or natural origin, and is one of the main sulfur compounds present in Earth’s atmosphere [11,12,13,14,15]. Volcanic eruptions and regular volcanic activities, such as fumaroles, constitute the main natural source in the atmosphere [16,17,18,19]. Biomass burning (e.g., wildfires, agricultural fires) also release SO₂ in the atmosphere, in addition to other sulfur compounds [20,21,22,23,24,25]. Anthropogenic emissions of SO₂ are mostly related to the burning of fossil fuels enriched in sulfur, and the effects of these outputs are heterogeneous across continents [26]. Prior to the implementation of new technologies and emission mitigation measures, a significant amount of SO₂ was linked to vehicular traffic; however, that output has now been significantly reduced [27,28,29,30]. Maritime shipping is one of the main anthropogenic sources of atmospheric SO₂, and affects vast areas of the planet [31,32,33]. Unlike CO, CO₂, and CH₄, which have atmospheric lifetimes in the order of months up to the scale of centuries, SO₂ is short-lived and—depending on the environment—may last in the atmosphere up to two days [34,35,36,37].

Black carbon (BC) is commonly released in the atmosphere following combustion processes [38,39,40,41], such as wildfires [42], and shows direct effects with respect to climate change mechanisms [43,44,45] and health hazards [46,47]. Like SO₂, BC is characterized by a short atmospheric lifetime, in the order of days [48,49]; atmospheric removal times of BC are, however, known to depend on the mixing state and morphology of particles [50,51,52]. Ports have also been linked to peaks in BC concentrations, with notable impacts on air quality [53,54,55].

The short atmospheric lifetimes of SO₂ and BC compared to other compounds subject to evaluation via the Proximity methodology should highlight notable differences between local and remote sources of emissions. In fact, CO₂ has a potential lifetime of centuries [56], while CH₄ is in the order of decades [57], and CO has a lifetime of approximately two months [58]. With lifetimes considerably longer than those of SO₂ and BC, it was possible to report differences between air mass aging categories [7].

This work is divided as follows: Section 2 describes the Lamezia Terme-LMT station, the instruments, datasets, and methodologies of the study; Section 3 shows the results of this evaluation; and Section 4 and Section 5 discuss the results and conclude the paper, respectively.

2. The LMT Station, Instruments, and Methods

2.1. The Lamezia Terme Regional Station in Calabria, Italy

The Lamezia Terme (code: LMT) is a regional/coastal observation site hosting instruments for the continuous measurements of greenhouse gases, aerosols, and meteorological parameters. It is located in the municipality of Lamezia Terme (province of Catanzaro), in the region of Calabria, Italy (Figure 1A). The station is located 600 m from the Tyrrhenian coast of the region (Lat: 38.8763° N; Lon: 16.2322° E; Elev: 6 m above sea level), in the westernmost sector of the Catanzaro Isthmus, which is the narrowest area of the entire Italian peninsula (≈32 km between the Tyrrhenian and Ionian coasts). The isthmus separates the Sila Massif in the north from the Serre Massif, in the south. Fully operated by CNR-ISAC (National Research Council of Italy–Institute of Atmospheric Sciences and Climate), the LMT site has been performing continuous measurements since 2015, although some measurements were already in place as early as 2014. The site is part of the World Meteorological Organization–Global Atmosphere Watch (WMO/GAW) network.

The peculiar orographic configuration of the isthmus, which is the result of tectonic activity and geodynamic processes at scales ranging from local to regional [59,60,61,62,63,64,65,66], is the main regulating factor of near-surface wind circulation. The Marcellinara Gap, located in the middle of the isthmus, channels winds through the area (Figure 1B); wind circulation is well oriented on a NE-ENE/W-WSW axis, and breeze regimes regulate local circulation [67,68]. Previous studies based on the results of short campaigns allowed for defining the pattern and behavior of vertical wind profiles in the area [69,70,71]. A more detailed study, based on a decade of measurements, has allowed for better understanding these patterns, as well as the transition from NE-ENE/W-WSW oriented near-surface winds and NW winds at higher altitudes, typical of large scale circulation in the area [72].

The LMT location in the Mediterranean, which is a well-known hotspot for climate change and air mass transport processes [73,74,75,76,77], exposes the observation site to Saharan dust events [78] and wildfire emissions from the same region [79] and other countries overlooking the Mediterranean Basin [80]. Local wind circulation is known, at the time of inversions between one regime and the other(s), to cause pollutants such as black carbon, subject to air mass transport at higher altitudes, to precipitate, thus resulting in peaks linked to changes in wind circulation [71,80].

Figure 1. (A) EMODnet DEM map [81] showing the main details of the LMT observation site’s location in the central Calabria region, and information on the known anthropogenic and natural sources of SO₂ in the sector. Additional details on the Aeolian Arc of volcanoes is available in previous research [9]. The “Messina Strait” label refers to the ports of Messina in Sicily, and Villa San Giovanni in Calabria. (B) Details of the Catanzaro Isthmus, with a highlight on LMT’s location in the Tyrrhenian coast and the main orographic features of the area.

Figure 1. (A) EMODnet DEM map [81] showing the main details of the LMT observation site’s location in the central Calabria region, and information on the known anthropogenic and natural sources of SO₂ in the sector. Additional details on the Aeolian Arc of volcanoes is available in previous research [9]. The “Messina Strait” label refers to the ports of Messina in Sicily, and Villa San Giovanni in Calabria. (B) Details of the Catanzaro Isthmus, with a highlight on LMT’s location in the Tyrrhenian coast and the main orographic features of the area.

LMT’s location also exposes the observation site to SO₂ emissions of anthropogenic and natural origin, such as maritime shipping and regular volcanic activity [9]. The Gioia Tauro port (52 km S-SW from LMT) is a major maritime shipping hub for cargo transport due to its intermediate location between the Suez Canal and the Gibraltar Strait: it is the busiest port in terms of cargo tonnage in southern Italy, and one of the busiest in the country [82]. The ports of Messina in Sicily and Villa San Giovanni in Calabria, located 90–95 km S-SW from LMT, are the busiest in Italy in terms of passenger traffic, as they connect mainland Italy with Sicily [82].

Several active volcanoes are located in the area: Mount Etna in Sicily [83,84,85,86,87] is located 160 km S-SW from LMT, while Stromboli [88,89] is the closest active volcano to LMT in the Aeolian Arc, 88 km in the W-SW direction. Vulcano, the southernmost Aeolian island, is a known source of SO₂ due to degassing via vents and fumarole [90]. A number of underwater volcanoes, both active and inactive, are part of the Aeolian Arc, including the Lametini twin volcanoes (LamN and LamS) named after the municipality of Lamezia Terme and located only 70 km W from the LMT observation site [9].

In addition to SO₂ emission sources on a regional scale, local sources have also been reported in past studies. Livestock farming has been suggested as a local CH₄ emission source, highlighted by the results of the O₃/NO_x methodology on LMT’s preliminary data [5]. This hypothesis was further corroborated using a longer data record (2015–2023) [9]. The observed black carbon has been attributed to wildfire emissions [79,80] and anthropogenic emissions [71,91].

At LMT, CH₄ was found to be characterized by peaks linked to the northeastern–continental wind sector of LMT, more exposed to anthropic influence (Figure 1B), especially during the winter season [92]. Specifically, a hyperbola branch pattern (HBP) was reported when comparing CH₄ mole fractions with wind speeds, as higher concentrations are linked to low wind speed, while low concentrations are linked to higher wind speeds. Surface O₃ at the site shows a completely different pattern, with western–seaside peaks linked to photochemical activity during the summer [6], and no evidence of a HBP. No evidence of such a pattern was reported for SO₂ [9], thus indicating that each compound is subject to variability patterns that depend on the balance between anthropogenic/natural emissions, location of emission sources compared to LMT, and air mass transport mechanisms.

2.2. Instruments, Datasets, and Methods

At LMT, SO₂ data in parts per billion (ppb) were gathered by a Thermo Scientific 32i (Franklin, MA, USA). The instrument’s principle of operation is based on ultraviolet (UV) light absorption by SO₂ and the consequent emission of light at specific wavelengths, up to their return to the normal, unexcited state. Ten measurements per minute were performed with a detection limit < 0.5 ppb of SO₂. More details on these measurements at LMT and Thermo 32i operations are available in previous research [9].

Equivalent black carbon (eBC) [93] measurements in micrograms per cubic meter (μg/m³) were performed at LMT by a Thermo Scientific 5012 MAAP (Multi-Angle Absorption Photometer) (Franklin, MA, USA) instrument. The principle of operation of the MAAP is based on the short-wave absorption of aerosols and BC in particular, with the consequent measurement of the sa (absorption coefficient) and eBC at 637 nanometers [94,95]. Data were gathered every minute and the minimum detection limit was <100 ng/m³. Additional details concerning MAAP measurements at the LMT are available in previous studies [79].

Positioned 10 m above sea level, a Vaisala WXT520 (Vantaa, Finland) measured the following meteorological data at the LMT: wind speed and direction, temperature, hail, air pressure, rain, and relative humidity. In this study, wind data were used, which were calculated by the instrument via ultrasonic transducers placed on a horizontal plane, and the measurement of deviations from regular traveling times of ultrasound pulses, caused by wind. Wind direction (WD) measurements had a precision of ±3 degrees, while wind speed (WS) measurements had a precision of ±0.3 m/s. These data were gathered on a per-minute basis, and additional information concerning WXT520 measurements at LMT is available in previous research [91,92].

The measurements of O₃ and NO_x used to define air mass aging and proximity categories were performed by Thermo Scientific 49i (Franklin, MA, USA) and Thermo Scientific 42i-TL (Franklin, MA, USA) instruments, respectively. Air mass aging and proximity categories were defined as follows, based on previous research [5,7]: with an O₃/NO_x ratio lower or equal to 10, measurements were attributed to the LOC (local) category; a ratio in the 10–50 range led to N–SRC (near source); with a ratio of 50–100, the measurement was attributed to the R–SRC (remote source) category; and if the O₃/NO_x ratio exceeded 100, the BKG (atmospheric background) category was attributed. The standard correction (“cor”) for aged air masses divides NO₂ concentrations by a factor of two [5]; the enhanced correction (“ecor”) introduced specifically at the LMT to reflects the observation site’s characteristics divides O₃ by a factor of two under specific circumstances, to account for photochemical peaks in the central Mediterranean during warm seasons [6,7]. Specifically, the correction was applied to measurements with a westerly wind direction (240–300° N) performed during diurnal hours (10:00–16:00 UTC) in spring and summer.

Additional details concerning the instruments performing O₃ and NO_x measurements, as well as their principles of operation are available in previous research [5,6,7,91].

All data (SO₂, eBC, WS, WD) were aggregated on an hourly basis to generate the averages used in this work. Seasonal categorizations were applied as follows: winter (JFD—January, February, December); spring (MAM—March, April, May); summer (JJA—June, July, August); and fall (SON—September, October, November).

Based on the characteristics of local wind circulation, the northeastern–continental (0–90° N) and western–seaside (240–300° N) sectors were considered.

In order to correlate measurements, the datasets were merged to generate subsets with valid data of two, or more, instruments. This resulted in a reduction in the total coverage rate, as shown in

Table 1. Coverage rates of the hourly SO₂ datasets used in this study. The SMTO dataset combines Thermo Scientific 32i (SO₂) and Vaisala WXT520 (meteorological parameters) measurements. SProx combines Thermo Scientific 32i measurements with hourly data for which a proximity category could be defined, i.e., valid Thermo Scientific 49i (O₃) and Thermo Scientific 42i-TL (NO_x) data. SMTOProx combines all previous datasets, thus resulting in lower coverage rates.

Year	Hours	SO2	Meteo	SMTO	SProx	SMTOProx
2016	8784	63.30%	96.34%	62.04%	61.28%	60.04%
2017	8760	87.76%	93.8%	83.86%	86.11%	82.23%
2018	8760	97.54%	77.05%	75.18%	97.04%	74.69%
2019	8760	80.67%	98.59%	80.65%	77.65%	77.63%
2020	8784	34.52%	99.98%	34.52%	33.03%	33.03%
2021	8760	39.08%	99.74%	39.07%	35.45%	35.44%
2022	8760	65.06%	89.85%	63.75%	62.48%	61.44%
2023	8760	48.59%	96.3%	47.37%	48.07%	46.86%
Total	70,128 1	64.56% 2	93.95% 2	60.80% 2	62.63% 2	58.92% 2

¹ Sum of all hours. ² Average rate.

(SO₂) and

Table 2. Coverage rates of the hourly eBC datasets used in this study. The BMTO dataset combines Thermo Scientific 5012 (eBC) and Vaisala WXT520 (meteorological parameters) measurements. BProx combines Thermo Scientific 5012 measurements with hourly data for which a proximity category could be defined, i.e., valid Thermo Scientific 49i (O₃) and Thermo Scientific 42i-TL (NO_x) data. BMTOProx combines all previous datasets, thus resulting in lower coverage rates.

Year	Hours	eBC	Meteo	BMTO	BProx	BMTOProx
2016	8784	93.75%	96.34%	93.06%	89.70%	89.20%
2017	8760	95.27%	93.8%	90.45%	92.24%	87.93%
2018	8760	95.61%	77.05%	73.61%	94.44%	72.5%
2019	8760	96.48%	98.59%	96.46%	92.26%	92.23%
2020	8784	96.61%	99.98%	96.60%	91.75%	91.74%
2021	8760	98.42%	99.74%	98.25%	77.39%	77.38%
2022	8760	97.43%	89.85%	88%	67.37%	65.98%
2023	8760	69.13%	96.3%	68.81%	60.51%	60.19%
Total	70,128 1	92.83% 2	93.95% 2	88.15% 2	83.20% 2	79.64% 2

¹ Sum of all hours. ² Average rate.

(eBC). In particular, any combination with the Proximity dataset was particularly susceptible to reductions in coverage, as the applicability of the methodology was dependent on multiple instruments operating at the same time.

All data were aggregated and processed in R 4.4.2 using the dplyr, ggpubr, ggplot2, zoo, tidyverse, and tseries libraries.

This study also introduced the Proximity Progression Factor (PPF) to compare the behavior of SO₂ and eBC concentrations in the progression from the LOC air mass aging category to BKG. The PPF was calculated as follows: LOC, N–SRC, R–SRC and BKG averaged concentrations of a given parameter, e.g., SO₂, were summed together; each concentration was divided by the sum to generate a percentage in the 0–1 range compared to the total; and three subtractions were calculated (LOC—N–SRC, N–SRC—R–SRC, R–SRC—BKG), with the PPF being the result of the algebraic sum of these three differences. Should “cor” (corrected) and “ecor” (enhanced corrected) concentrations for R–SRC and BKG be used, the results would be the PPF_c and PPF_ec, respectively. A flow chart showing the main steps of PPF’s calculation is shown in Figure 2.

3. Results

3.1. Concentration Variability and Proximity Progression Factor (PPF)

The average concentrations of SO₂ and eBC were calculated based on proximity categories and wind corridors, and the results are shown in

Table 3. Concentrations of SO₂ (ppb) and eBC (µg/m³) on a per-category basis, also including the ±1σ interval (one standard deviation). Data are divided by wind sector. The “cor” and “ecor” suffixes refer to the standard and enhanced corrector factors of the O₃/NO_x ratio.

Category	SO2 (ppb)			eBC (µg/m3)
	All	Nor. East	West	All	Nor. East	West
LOC	0.145 ±0.280	0.125 ±0.257	0.237 ±0.399	0.879 ±0.632	0.887 ±0.644	0.715 ±0.517
N–SRC	0.216 ±0.357	0.174 ±0.319	0.243 ±0.381	0.369 ±0.341	0.404 ±0.245	0.334 ±0.295
R–SRC	0.222 ±0.402	0.176 ±0.351	0.241 ±0.424	0.273 ±0.605	0.229 ±0.119	0.270 ±0.581
BKG	0.178 ±0.309	0.060 ±0.083	0.201 ±0.334	0.264 ±0.788	0.198 ±0.089	0.272 ±0.846
R–SRCcor	0.255 ±0.449	0.196 ±0.365	0.269 ±0.465	0.291 ±0.730	0.259 ±0.130	0.290 ±0.785
BKGcor	0.206 ±0.376	0.156 ±0.329	0.227 ±0.400	0.266 ±0.611	0.216 ±0.111	0.265 ±0.590
R–SRCecor	0.256 ±0.432	0.196 ±0.365	0.264 ±0.438	0.294 ±0.637	0.259 ±0.130	0.294 ±0.653
BKGecor	0.176 ±0.323	0.156 ±0.329	0.194 ±0.340	0.253 ±0.657	0.216 ±0.111	0.245 ±0.649

and Figure 3. From this evaluation, it is possible to infer that SO₂ did not behave like eBC and other compounds previously subject to a similar analysis (CO, CO₂, CH₄) [7]. In fact, LOC yielded lower concentrations compared to N–SRC and R–SRC. Conversely, eBC showed a pattern consistent with the previously analyzed compounds, with a clear progression from higher concentrations in LOC to lower concentrations in BKG.

Seasonal variability has also been considered for SO₂ (Figure 4) and eBC (Figure 5). In the case of SO₂, these plots show that N–SRC and R–SRC generally yielded concentrations higher than those of LOC. In the case of eBC, although the difference between wind sectors showed degrees of variability across seasons, the progression from LOC to BKG was very well-defined.

The statistical significance of the differences between air mass aging categories was assessed. SO₂ and eBC concentrations were tested for normality using the Shapiro–Wilk [96] and Jarque–Bera [97] tests; both parameters were shown not to have a normal distribution with a very high level of significance (p-values < 2.2 × 10⁻¹⁶). Consequently, the Kruskal–Wallis [98] method was used to verify the significance of the differences between standard and corrected categories, which also yielded very significant results (p-values < 2.2 × 10⁻¹⁶). The pairwise Wilcoxon (Mann–Whitney U) test [99,100] with Bonferroni correction [101,102] was also used to assess the significance of the differences between pairs of main proximity categories, also yielding statistically significant results (p-values <<< 0.05).

Considering SO₂’s anomalous behavior compared to CO, CO₂, CH₄, and eBC, the Proximity Progression Factor (PPF) introduced in this study was used to assess the variability of each parameter and determine the extent of SO₂’s pattern. The results, shown in

Table 4. Proximity Progression Factor (PPF) of SO₂ and eBC, from this study, compared with the PPF of CO, CO₂, and CH₄ based on the findings of previous research on concentration variability by air mass aging category. PPF_c and PPF_ec consider “cor” and “ecor” the corrections of R–SRC (remote source) and BKG (atmospheric background) categories.

Parameter	PPF	PPFc	PPFec
SO2	−0.043	−0.074	−0.039
eBC	0.345	0.340	0.349
CO	0.130	0.123	0.124
CO2	0.024	0.022	0.022
CH4	0.024	0.023	0.023

, indicate a consistent negative PPF for SO₂, which was not reported for the other parameters. Concentration data of CO, CO₂, and CH₄ were previously calculated in another work [7].

3.2. Standard and Seasonal Daily Cycles

LMT observations were largely influenced by daily cycle variability, which is the result of local wind circulation. Changes between northeastern–continental and western–seaside winds occurred during “inversions”, which can result in peaks in pollutant concentrations [5,71,91]. The daily cycles of SO₂ and eBC are shown in Figure 6, and indicate, for eBC in particular, increased LOC concentrations occurring in early morning hours and in the late afternoon, a pattern consistent with the findings of previous studies [5,71,91].

The daily cycles were further divided on a seasonal basis, with the results shown in Figure 7 (winter), Figure 8 (spring), Figure 9 (summer), and Figure 10 (fall). From these plots, it is possible to infer that the extent of inversion patterns’ influences over observed concentrations of eBC in particular reflects seasonal changes.

3.3. Analysis with Wind Direction and Speed

In addition to the analysis of concentration variability based on wind corridors, polar plots were generated to assess the behavior of SO₂ (Figure 11) and eBC (Figure 12) based on proximity categories. This analysis is based on previous research which allowed for verifying the correlation between each proximity category and wind [7].

A previous study on CH₄ showed the occurrence of a HBP, with low concentrations generally linked to high speeds and, vice versa, high concentrations linked to low wind speeds [92]. This pattern was not reported for O₃ [6] and SO₂ [9] at the site. Using proximity categories, it was demonstrated that the high concentration–low wind speed combination in CH₄ was mostly restricted to LOC [7].

In this study, SO₂ (Figure 13) and eBC (Figure 14) were evaluated based on concentration–wind speed variability based on proximity categories. The same pattern previously reported for CH₄ can be seen for eBC, as LOC is linked to low wind speeds and high eBC concentrations.

3.4. Analysis of Weekly Cycles

At the LMT, several studies assessed the weekly (MON-SUN) variability of parameters to verify the possible anthropogenic influences [6,7,9,92,103]. For SO₂, no notable weekly patterns were reported; however, previous evaluations did not consider proximity categories and possible enhanced influences of anthropogenic emissions over the LOC category, more susceptible to short term responses. The weekly patterns of SO₂ (Figure 15) and eBC (Figure 16) accounting for proximity categories are hereby reported; in the case of eBC, differences between weekdays reflect the differences in absolute concentrations between proximity categories, while in the case of SO₂, the reported concentrations tend to overlap and no specific pattern is reported except for a northeastern reduction of BKG from MON to THU, followed by an increase in concentrations up to SUN. These plots also consider both the “cor” and “ecor” correction factors, with the latter being limited to warm seasons (March–August) [7].

For a more detailed understanding of the possible weekly behaviors in observed data, the results were evaluated by comparing weekday (WD, MON-FRI) concentrations with their WE (weekend, SAT-SUN) counterparts using Kruskal–Wallis tests [98], based on proximity categories and wind corridors. The results are shown in

Table 5. Results of the Kruskal–Wallis tests (p-values) aimed at verifying the statistical significance of the difference between average WD (weekday, MON-FRI) and WE (weekend, SAT-SUN) concentrations for each proximity category and wind corridor. Values lower than 0.05 indicate a statistically significant weekly cycle.

Category	SO2 (ppb)			eBC (µg/m3)
	All	Nor. East	West	All	Nor. East	West
LOC	<0.05	<0.05	0.15	<0.05	<0.05	<0.05
N–SRC	0.72	<0.05	0.15	<0.05	0.51	<0.05
R–SRC	0.90	0.62	0.64	<0.05	<0.05	<0.05
BKG	0.11	0.93	<0.05	0.92	0.08	0.94
R–SRCcor	0.91	0.62	0.77	0.56	0.45	0.77
BKGcor	0.49	0.84	0.17	<0.05	0.07	<0.05
R–SRCecor	0.30	0.62	0.27	<0.05	0.45	<0.05
BKGecor	0.82	0.84	0.48	<0.05	0.07	0.06

.

Although the results allow for pinpointing the differences between categories and wind sectors in terms of weekly cycles, the presence of multiple sources (both natural and anthropogenic), combined with seasonal patterns in anthropic activities, requires the results to be assessed on a seasonal basis. Therefore, seasonal analyses have been performed for each season: winter (

Table 6. Results of the Kruskal–Wallis tests (p-values) aimed at verifying the statistical significance of the difference between average WD (weekday, MON-FRI) and WE (weekend, SAT-SUN) concentrations for each proximity category and wind corridor, during the winter season. Values lower than 0.05 indicate a statistically significant weekly cycle. “>0.05” values indicate results slightly higher than the required threshold of significance, which do not meet the criteria to pass the test.

Category (Winter)	SO2 (ppb)			eBC (µg/m3)
	All	Nor. East	West	All	Nor. East	West
LOC	<0.05	<0.05	0.65	0.81	0.37	0.34
N–SRC	0.86	0.29	0.08	<0.05	0.15	<0.05
R–SRC	<0.05	0.65	0.35	<0.05	0.27	0.25
BKG	>0.05	0.23	0.42	0.78	>0.05	0.97
R–SRCcor	0.75	0.88	0.68	0.61	0.76	0.17
BKGcor	0.13	0.59	0.38	<0.05	<0.05	<0.05
R–SRCecor	0.75	0.88	0.68	0.61	0.76	0.17
BKGecor	0.13	0.59	0.38	<0.05	<0.05	<0.05

), spring (

Table 7. Results of the Kruskal–Wallis tests (p-values) aimed at verifying the statistical significance of the difference between average WD (weekday, MON-FRI) and WE (weekend, SAT-SUN) concentrations for each proximity category and wind corridor, during the spring season. Values lower than 0.05 indicate a statistically significant weekly cycle. “>0.05” values indicate results slightly higher than the required threshold of significance, which do not meet the criteria to pass the test.

Category (Spring)	SO2 (ppb)			eBC (µg/m3)
	All	Nor. East	West	All	Nor. East	West
LOC	0.45	0.09	0.93	0.62	0.25	0.17
N–SRC	0.25	<0.05	0.50	<0.05	0.12	<0.05
R–SRC	<0.05	<0.05	0.36	<0.05	0.14	<0.05
BKG	<0.05	N/A	<0.05	<0.05	0.31	<0.05
R–SRCcor	0.84	0.14	0.60	0.74	0.53	0.72
BKGcor	<0.05	<0.05	<0.05	<0.05	0.28	<0.05
R–SRCecor	<0.05	0.14	<0.05	<0.05	0.53	<0.05
BKGecor	>0.05	<0.05	0.94	<0.05	0.28	<0.05

), summer (

Table 8. Results of the Kruskal–Wallis tests (p-values) aimed at verifying the statistical significance of the difference between average WD (weekday, MON-FRI) and WE (weekend, SAT-SUN) concentrations for each proximity category and wind corridor, during the summer season. Values lower than 0.05 indicate a statistically significant weekly cycle. “>0.05” values indicate results slightly higher than the required threshold of significance, which do not meet the criteria to pass the test.

Category (Summer)	SO2 (ppb)			eBC (µg/m3)
	All	Nor. East	West	All	Nor. East	West
LOC	<0.05	>0.05	<0.05	<0.05	0.29	0.06
N–SRC	<0.05	<0.05	<0.05	<0.05	0.85	>0.05
R–SRC	0.70	0.17	0.95	<0.05	0.14	<0.05
BKG	<0.05	N/A	<0.05	0.26	N/A	0.18
R–SRCcor	0.54	N/A	0.21	0.78	N/A	0.87
BKGcor	0.86	0.28	0.72	0.33	0.75	0.34
R–SRCecor	0.91	N/A	0.84	>0.05	N/A	>0.05
BKGecor	0.87	0.28	0.90	0.78	0.75	0.50

), and fall (

Table 9. Results of the Kruskal–Wallis tests (p-values) aimed at verifying the statistical significance of the difference between average WD (weekday, MON-FRI) and WE (weekend, SAT-SUN) concentrations for each proximity category and wind corridor, during the fall season. Values lower than 0.05 indicate a statistically significant weekly cycle.

Category (Fall)	SO2 (ppb)			eBC (µg/m3)
	All	Nor. East	West	All	Nor. East	West
LOC	<0.05	<0.05	0.55	<0.05	<0.05	<0.05
N–SRC	<0.05	<0.05	0.43	<0.05	0.23	<0.05
R–SRC	0.29	0.12	0.12	0.86	<0.05	0.36
BKG	0.57	0.46	0.82	0.65	<0.05	0.90
R–SRCcor	0.87	0.61	0.44	0.62	0.44	0.54
BKGcor	0.50	0.18	0.47	0.98	<0.05	0.32
R–SRCecor	0.87	0.61	0.44	0.62	0.44	0.54
BKGecor	0.50	0.18	0.47	0.98	<0.05	0.32

).

From these results, it is possible to infer that changes in the significance of weekly cycles between season are representative of alternating sources of emissions which depend on seasonality/temperatures, i.e., domestic heating.

4. Discussion

At the Lamezia Terme WMO/GAW observation site in Calabria, Italy, air mass aging and proximity categories based on the O₃/NO_x ratio were first applied to preliminary data from the site [5] and later extended to nine years (2015–2023) of CO, CO₂, and CH₄ data at the site [7], also accounting for new correction factors based on local O₃ behavior [6]. Prior to this study, this methodology has never been applied to SO₂ and, more importantly, to an aerosol such as eBC. In the literature, studies have assessed BC variability for presumably aged air masses using other methods [104,105]. An earlier study from the Italian network found evidence of correlations between BC and air mass aging categories; however, it was limited to a short summertime observation period [10].

Correction factors were applied to this method for a number of reasons: the first correction, which is instrumental in nature, was due to the challenge of measuring “true NO_x” concentrations, as instruments relying on heated molybdenum converts have been reported to overestimate NO₂ in aged air masses [8]. Several works have described interferences of chemical and physical factors over NO_x measurements [106,107,108,109,110,111,112,113]. Another reason for the implementation of correction factors is related to the strict requirements of the BKG category: a previous work reported that only 0.89% (2017), 0.26% (2018), 0.07% (2019), and 0.95% (2022) of all measurements met the criteria for BKG at LMT, while the implementation of BKG_cor and BKG_ecor significantly increased those figures (e.g., 8.15% for BKG_cor and 5.46% for BKG_ecor in 2019) [7]. Changes in the distribution of data were also reported: BKG, for instance, showed a blind spot between 30 and 60° N, which is in the direction of Lamezia Terme’s downtown area; at 60° N, wind channeled through the Marcellinara Gap (Figure 1B) pointed directing at LMT and, under exceptional conditions, was unperturbed enough from anthropogenic emissions to meet the BKG criteria. Conversely, the BKG_cor and BKG_ecor categories also included, in the same study, measurements in the 30–60° N due to less strict conditions.

The area where LMT is located is characterized by a dual rural/urban nature: livestock and agricultural emissions have been reported in previous research [5,92], in addition to emissions from the transport sector (highways, railways, aviation) [5,92], and biomass burning related to domestic heating during cold seasons [103] and open fire phenomena during warm seasons [79,80]. This behavior was also reported in the evaluation of the OWE (Ozone Weekend Effect) [114,115,116,117,118], which resulted in characteristics classified as intermediate between rural and urban areas [6]. For this reason, source apportionment and atmospheric tracers were required for a better understanding of the balance between emission sources.

In detail, SO₂ and eBC contributions to LMT’s measurements varied in extent and nature: a previous study found substantial evidence of volcanic activity as responsible for natural SO₂ outputs in the western sector of LMT (Figure 1A), due to the presence of multiple active volcanoes within 120 km from the observation site [9]. Moreso, the Gioia Tauro port, was deemed responsible for anthropogenic emissions in the same area. This study also considered the Messina Strait system, with the ports of Messina in Sicily and Villa San Giovanni in Calabria constituting the busiest maritime passenger traffic in the entire country [82] (Figure 1A). The sources of eBC were more various, as LMT is exposed to regional [79] and Mediterranean [80] open fire emissions, and biomass burning [103].

The station’s location in central Calabria was also a main driver of observations, as the Catanzaro isthmus—the narrowest point in Italy—separates the Tyrrhenian (west) and Ionian (east) seas, as well as the mountain ranges Sila (north) and Serre (south), thus resulting in a unique configuration (Figure 1B). The peculiar wind patterns resulting from this configuration [67,68,72] had a relevant influence on the observed concentrations of gases and aerosols [6,7,80,91,92], including SO₂ [9].

As reported in a previous study [7], this methodology was susceptible to coverage rate losses as it required multiple instruments to operate at the same time: in this work, SO₂ and eBC concentrations with a defined proximity category required three instruments each (Table 1 and Table 2). With the implementation of wind speed and direction, a total of four instruments were required, thus resulting in combined coverage rates of 58.92% in terms of hourly data for SO₂, and a 79.64% coverage rate for eBC.

With multiple parameters now subject to evaluation under the Proximity method, a mean of comparison between their behaviors was required. This study introduced the Proximity Progression Factor (PPF) (Figure 2) to provide a tool by which the progression from LOC to BKG (and its corrected counterparts) is representative of a transition from enriched local emissions to lower concentrations typical of the atmospheric background. The first step toward the implementation of this method was calculating average concentrations, on a per-category basis, of SO₂ and eBC (Table 3, Figure 3). These averages underlined the anomalous behavior of SO₂ compared to eBC itself, as well as CO, CO₂, and CH₄, as previously described [7]: in fact, while eBC consistently transitioned from high LOC to lower BKG concentrations, SO₂ tended to peak at N–SRC and R–SRC. The introduction of the PPF in this work constitutes an attempt to provide researchers, and possibly policy makers/regulators, with a tool which can help defining tendencies in the variabilities of pollutants and lead to a more accurate understanding of the balances between natural and anthropogenic sources of emission.

When seasons and wind corridors were considered, SO₂ (Figure 4) yielded generally higher concentrations from the western sector, which is consistent with previous findings on volcanic and maritime shipping emissions [9]. N–SRC and R–SRC both tended to peak during all seasons from the continental sector; however, during winter and fall, R–SRC yielded lower concentrations compared to N–SRC and LOC, thus indicating a shift that may be attributable to changes in maritime emissions, such as reduced passenger traffic during cold seasons compared to warm seasons, as operations were increased due to tourism [82].

The seasonal behavior of eBC (Figure 5) was more consistent with the patterns normally reported for parameters other than SO₂ [7]. However, the differences between continental and seaside concentrations varied depending on the season and peak during summer, when biomass burning on the continent was reduced [103]; however, open-fire emissions tended to increase [79,80]. Furthermore, during the summer, seaside LOC concentrations exceeded their continental counterparts, thus confirming the shift in the nature of eBC releases in the atmosphere. The statistical methodologies used to evaluate the results shown in Figure 4 and Figure 5 show that the differences between proximity categories were relevant.

With the variability in concentrations between proximity categories, the PPF allowed for introducing an objective tool to assess SO₂’s anomaly. It was, in fact, the only parameter to yield a negative PPF at −0.043 (with corrections at −0.074 and −0.039), while eBC’s value was the highest reported among all parameters at 0.345 (with corrections in the 0.340–0.349 range). CO yielded a range of 0.123–0.130, higher than that of CO₂ (0.022–0.024) and CH₄ (0.023–0.024). Considering that the atmospheric lifetimes of SO₂ [34,35,36,37] and eBC [48,49] were very short, the reported differences in terms of PPF could not be explained by the atmospheric lifetime, and should be representative of a local-to-remote difference between SO₂ and eBC sources in the southern Italian peninsula.

Ever since their introduction [3,4], proximity categories have been affected by a lack of accurate spatial resolution, which was also highlighted in previous research on LMT’s observations [7]. SO₂’s anomaly, however, combined with the punctual known emission sources located in the Aeolian Arc, Sicily, and Calabria could, therefore, be exploited to add—although uncertain and not accurately defined—estimates on spatial resolution ranges and thresholds to the main proximity categories, limited to the western sector. LOC would be limited to a ≈70 km radius from LMT in the western direction, as that is lower than the distance to the first known source of natural SO₂, i.e., the Stromboli volcano. BKG would be representative of air masses originating 200 km or more in that direction, with no tangible effects by volcanoes and maritime shipping routes to and from the Gioia Tauro port. Consequently, N–SRC and R–SRC would fill the intermediate niche in the ≈70–200 km range; however, their exact boundaries cannot be presently determined.

These ranges constitute a first attempt at providing spatial resolution to the method based on known sources of emission, and SO₂’s anomalous PPF, and need to be further verified in future studies via the implementation of additional methodologies.

Changes in PPF variability, i.e., a negative PPF differing from the standard values observed for all parameters with the exception of SO₂, could potentially be used by policymakers and regulators to assess regional scale influences over a given area. Standard (positive) PPF values would be expectable in areas with punctual emission sources, while negative or even neutral values would indicate contributions on a regional scale that affect air quality at multiple levels, and would therefore require additional efforts be made to mitigate risks for human health and the environment.

In addition to new hypotheses on source apportionment on a regional scale, the local behavior can also be assessed in greater detail. The daily cycle, which results in peculiar patterns of gases and aerosols at LMT [6,9,71,80,91,92], was hereby assessed for SO₂ and eBC based on proximity categories (Figure 6). From the main daily cycle, it is possible to infer that LOC eBC was very susceptible to the local wind inversion patterns affecting the area (early morning and late afternoon), during which wind directions changed and speeds were affected by the inversion itself, thus leading to increases in the near-surface concentration of pollutants that would normally be subject to air mass transport at higher altitudes [71,91]. During the winter (Figure 7), this phenomenon was amplified for eBC, while SO₂ was affected by an increase in the late morning hours and higher degrees of variability between proximity categories, with many concentrations overlapping. BKG, however, was consistently lower than other categories, thus representing an unaffected atmospheric background concentration, with no natural or anthropogenic emissions influencing SO₂ mole fractions. During the spring season (Figure 8), although BKG yielded low concentrations for both SO₂ and eBC, notable gaps were reported between 04:00 and 08:00 UTC, with no available data falling under the BKG category throughout the entire observation period (2016–2023). In fact, BKG’s strict requirements, already reported for LMT’s preliminary data [5], have led to the introduction of “cor” and “ecor” corrections [7], which can result in higher SO₂ and eBC concentrations. Despite these differences, the general pattern of eBC (early morning and late afternoon inversion peaks) and SO₂ (late morning peaks) were clearly present during this season. During the summer (Figure 9), eBC’s LOC concentration during diurnal hours tended to match those of the other categories other than BKG, thus confirming the shift in emission sources that occurred between seasons; in fact, the peaks linked to wind inversion were consistent with the transport of open-fire emissions [79,80]. Conversely, SO₂ showed a peak in N–SRC between 10:00 and 15:00 UTC not seen in other seasons, and possibly linked to maritime passenger traffic. When the fall season (Figure 10) is considered, the two parameters showed distinct behaviors, with SO₂ reporting multiple overlapped between categories, although the early morning inversion peak of LOC was clearly present, while eBC showed a pattern closer to its winter counterpart, once again indicating seasonal changes between emission sources.

These evaluations showed differences between the main R–SRC and BKG categories, and their corrected (cor, ecor) counterparts; these differences are caused by the criteria required for each category/correction, and notable changes in the number of data being considered when a correction is introduced. These differences can be clearly seen in the polar plots of SO₂ (Figure 11) and eBC (Figure 12), where LOC and N–SRC dominate in terms of number of hourly data [7], while the introduction of correction factors affects in particular the coverage of BKG. Specifically, the standard BKG category, which has yielded generally low concentrations, is represented by a very limited number of data, while BKG_cor and BKG_ecor account a higher number of observations due to the correction factors applied to NO₂ and O₃. Specifically, “cor” concentrations tended to increase the number of hours falling under BKG, due to NO₂ mole fractions being multiplied by a factor of 0.5 and the consequent increase in O₃/NO_x ratios; “ecor” compensated for that by applying a 0.5 factor to O₃ under specific conditions, thus reducing the number of hours falling under BKG. These shifts constitute one of the main limitations of the Proximity method, as instruments measuring NO_x relying on heated molybdenum converters can lead to substantial uncertainties in NO₂ measurements, which in turn affect the O₃/NO_x ratio. The future implementation of instruments not affected by NO₂ measurement uncertainties would significantly reduce the need to rely on correction factors.

R–SRC, R–SRC_cor, R–SRC_ecor, BKG_cor, and BKG_ecor plots of SO₂ show high concentrations from the west, a pattern compatible with natural and anthropogenic emissions from that sector, which are nearly absent in LOC. These differences were not seen in eBC, where all categories past N–SRC showed very similar patterns and LOC showed very high concentrations from all sectors, specifically from the northeastern corridor which is scarcely reported for SO₂. It is worth mentioning that BKG categories showed the same blind spot at 30–60° N (downtown Lamezia Terme), while at 60° N—which, as reported before, was in the direction of the Marcellinara Gap (Figure 1B)—a number of measurements were present. This orientation may be compatible with a possible Ionian source of these concentrations, channeled through the Marcellinara Gap and less exposed to anthropogenic influences.

The variability of proximity categories with wind is further described in Figure 13 (SO₂) and Figure 14 (eBC), which show the behavior compared to wind speeds from multiple sectors. While other pollutants tended to peak from the northeast [92], SO₂ showed prevailing concentrations from the west, thus providing further evidence of natural and anthropogenic sources of emissions from that sector [9]. Conversely, eBC’s behavior was very closely related to that of CH₄ and similar parameters [7], with northeastern peaks in concentration linked to low wind speeds that resulted in a HBP.

Additional details on source apportionment at LMT frequently resulted from weekly evaluations [103], as daily, seasonal, and annual cycles may have been influenced by natural and anthropogenic emissions alike, while weekly patterns were purely anthropogenic in nature [114,116,119,120,121]. The general weekly cycle of SO₂ (Figure 15) did not show particular shifts except for reduced BKG concentrations on Thursday from the northeastern sector, and had general weekly trends with peaks on Monday and Sunday. Assuming this pattern is the result of anthropogenic emissions, the behavior is anomalous as anthropogenic emissions would build up during the course of a standard week; this pattern requires further investigation, possibly relying on the implementation of additional atmospheric tracers. The cycle of eBC (Figure 16) showed variations mostly driven by differences between categories in terms of absolute concentrations, thus reflecting a combination of emission sources.

A more detailed understanding of weekly cycles was possible by comparing averaged weekday (WD, MON-FRI) with weekend (WE, SAT-SUN) concentrations using the Kruskal–Wallis test [98] and assessing the statistical significance of the differences between WD and WE, also accounting for wind sectors (Table 5). Due to their proximity, LOC and N–SRC should be more affected by weekly cycles especially from the northeastern–continental sector of LMT, while N–SRC and—in particular—BKG should be less affected due to contributions by remote sources. SO₂ and eBC showed two distinct behaviors: with the exception of the western BKG sector of SO₂, no statistically significant weekly cycle was present in remote proximity categories, while close categories did yield statistically significant values from the northeast, thus providing evidence of contributions from fossil fuel burning. eBC showed a higher number of weekly cycles, indicating a significant contribution from combustion processes (fossil fuel burning, domestic heating), which is consistent with previous findings concerning the weekly behavior observed at the LMT [103].

Emission sources can change based on seasonality: during the winter season, eBC peaks were attributed to domestic heating and increased anthropogenic emissions, while the summer season had no domestic heating but was widely affected by wildfire emissions at various scales [79,80]. For this reason, seasonal weekly cycles were also assessed (Table 6, Table 7, Table 8 and Table 9). The highest number of significant weekly cycles was observed during spring (Table 7) for SO₂ and eBC alike, a season affected by intermediate temperatures between winter and summer, and the coexistence of multiple sources of emission. The fall season (Table 9) indicated, for both parameters, a prominence of LOC and N–SRC weekly cycles, which is consistent with anthropic activities nearby. The two opposite seasons in terms of known sources and temperatures, winter (Table 6) and summer (Table 8) also showed peculiar patterns. Summertime weekly cycles from the western sector across multiple proximity categories may indicate peaks in maritime shipping and passenger transport, while—during the winter—only LOC resulted in a significant weekly cycle, but from the northeast. eBC’s lack of prominent summertime weekly cycles, especially from the northeastern sector, was consistent with increased wildfire outputs which did not follow a weekly cycle and occurred randomly during the course of a standard week. During the winter, the absence of significant eBC from the northeastern sector was consistent with the balance between fossil fuel burning and domestic heating, which alternate their peaks during a standard week.

Overall, it is possible to infer that the Proximity methodology can be applied to gases such as SO₂ and can also be applied to aerosols such as eBC, thus underlining the potential of this method as an additional tool toward source apportionment and the differentiation of natural and anthropogenic sources. Although SO₂’s anomaly has allowed for assessing, for the first time, the spatial resolution of proximity categories, future research accounting for additional atmospheric tracers would be required to improve the accuracy of the method and enhance its application in the field of air quality monitoring [122,123,124,125].

One possibility would be using stable carbon isotopes in CO₂, paired with SO₂ measurements, to differentiate volcanic emissions from their anthropogenic counterparts [126,127,128,129], which are characterized by a different isotopic fingerprint. The LMT is part of the developing cross-country network of atmospheric stations performing continuous measurements of stable carbon isotopes in CO₂ and CH₄, and the integration of these measurements with SO₂ evaluation would significantly improve the potential of the Proximity method.

5. Conclusions

Using eight years (2016–2023) of continuous measurements at the Lamezia Terme (code: LMT) observation site in Calabria, Southern Italy, sulfur dioxide (SO₂) and equivalent Black Carbon (eBC) data were assessed, for the first time, using the ozone-to-nitrogen oxide ratio (O₃/NO_x) “Proximity” methodology. The analysis has shown two peculiar behaviors, with eBC reflecting a regular progression from the LOC (local) category, characterized by higher concentrations, to the BKG (atmospheric background) category, with lower concentrations. This pattern is consistent with that observed in a previous study for carbon monoxide (CO), carbon dioxide (CO₂), and methane (CH₄) at the same observation site. SO₂, however, has shown a unique behavior, with N–SRC (near source) and R–SRC (remote source) generally yielding higher concentrations than LOC and BKG. This anomaly was first tested and evaluated using the Proximity Progression Factor (PPF) introduced in this study to assess the patterns of observed parameters and determine their behavior; using the PPF, SO₂ yielded a negative value, while eBC, CO, CO₂, and CH₄ all yielded variable, but positive, values. Via the PPF, future research based on the Proximity methodology would be able to provide a single and consistent assessment on the variability of gases and aerosols based on air mass aging categories.

By exploiting SO₂’s anomaly, in this study, an attempt was made to implement spatial resolution to the Proximity method, which is affected by qualitative categories based on air mass aging and photochemistry. By analyzing known SO₂ sources of emission on a regional scale, an approximate range of ≈70–200 km from LMT was defined for N–SRC and R–SRC, thus constituting the first known attempt to associate air mass aging categories with a distance. These hypotheses need to be confirmed by future works relying on air mass transport modeling, and additional atmospheric tracers.

The analysis of daily, seasonal, and weekly patterns has allowed for observing changes in concentrations influenced by local near-surface wind circulation at the LMT, which is a key driver of air mass transport in the area. The study highlights the potential of the Proximity method in source apportionment efforts; however, future studies need to rely on additional atmospheric tracers to improve the spatial resolution of the method and better discriminate between emission sources.