Urban Precipitation Scavenging and Meteorological Influences on BTEX Concentrations: Implications for Environmental Quality

Abstract

This study provides an assessment of BTEX compounds—benzene, toluene, ethylbenzene, and xylene isomers—in urban precipitation collected in the city of Novi Sad, Republic of Serbia, during autumn and winter 2024, analyzed by gas chromatography-mass spectrometry (GC-MS). By combining chemical analysis with meteorological observations and HYSPLIT backward trajectory modeling, the study considers the mechanisms of BTEX removal from the atmosphere via wet scavenging and highlights the role of local weather conditions and long-range atmospheric transport in pollutant concentrations. During the early observation period (September to late November), average concentrations were 0.45 µg/L benzene, 3.45 µg/L ethylbenzene, 4.0 µg/L p-xylene, 2.31 µg/L o-xylene, and 1.32 µg/L toluene. These values sharply dropped to near-zero levels in December for benzene, ethylbenzene, and xylenes, while toluene persisted at 1.12 µg/L. A pronounced toluene spike exceeding 6 µg/L on 28 November was likely driven by transboundary air mass transport from Central Europe, as confirmed by trajectory modeling. The environmental risks posed by BTEX deposition, especially from toluene and xylenes, underline the need for regulatory frameworks to include precipitation as a pathway for pollutant deposition. It should be clarified that the identified risk primarily concerns aquatic organisms, due to the potential for BTEX infiltration into surface waters and subsequent ecotoxicological impacts. Incorporating such monitoring into EU policies can improve protection of air, water, and ecosystems.

1. Introduction

Benzene, toluene, ethylbenzene, and xylene isomers (BTEX) are known environmental pollutants [1,2] that belong to a group of volatile organic compounds (VOCs). These compounds are typically associated with petroleum-based products, including refining, industrial emissions, and vehicular exhaust. BTEX compounds are precursors of the main photochemical tropospheric ozone, secondary organic aerosols [3], and hazardous air pollutants, since chronic toxicity can be induced even at small concentrations [4]. According to the data from the International Agency for Research on Cancer, benzene should be classified as a high-priority public health hazardous agent and Group I carcinogen; ethylbenzene would be classified as a Group IIB suspected carcinogenic agent; and both toluene and the isomers of xylene are placed in Group III based on their neurotoxicity effect [5,6].

As human activities are more pronounced in cities, urban atmosphere, water, and soil are particularly susceptible to BTEX pollution. BTEX levels in various forms of atmospheric water can be influenced by gas-phase concentrations in ambient air, water solubility, Henry’s law constants, frequency and intensity of precipitation events, gas–water interactions [7,8], and origin of air masses [9]. Atmospheric BTEX compounds can partition into the gas phase and be deposited via atmospheric water droplets or adsorbed onto particles [10]. A study by Bąk et al. [11] showed that BTEX concentrations in precipitation in urban environments depend on climatic conditions and emissions from various anthropogenic sources. A study by Šoštarić et al. [12] also showed that gas mixing ratios, precipitation characteristics, and meteorology affect BTEX distribution in Belgrade, Serbia. On the other hand, a study by Mullaugh et al. [13] showed that there was no significant impact of air mass back-trajectory on the concentrations of BTEX observed in rain Wilmington, NC, USA. BTEX concentrations in rain samples, as presented by Okochi et al. [14], showed values higher than those estimated from Henry’s law. It now appears that atmospheric precipitation might contribute far more towards the removal of BTEX species in the ambient atmosphere compared to their previously assumed consumption rates. These authors found that all surface-active agents in rain droplets were responsible for observable supersaturation, while rainfall intensities were regarded as being of no significance. Measurements taken by Liu et al. [9] indicated the rather negligible role of wet deposition as an efficient process for atmospheric BTEX removal. The major atmospheric sink would, therefore, still involve light-mediated reactions with a hydroxide group (OH)· or nitrogen radicals. Though atmospheric water is crucial for the biogeochemical cycling of these substances, there are still facts about their behavior and destiny that need to be studied scientifically [15]. Understanding the occurrence and distribution of BTEX in precipitation is crucial for environmental protection and human health, as once deposited into the environment, these compounds can infiltrate soils and surface waters, posing risks to both terrestrial and aquatic ecosystems [16,17]. Since BTEX compounds are soluble in water and migrate into it, BTEX compounds present in precipitation mainly end up in soil, open water, and groundwater [18,19]. Their toxicity, involving the possibility of endocrine disruption, presents a great risk to human health and ecosystems [17,20], as well as an increased risk of cancer [21]. It is important to determine risk assessment scores for those species that are threatened, whether they are humans or aquatic organisms [1]. Therefore, studying BTEX concentrations is vital for evaluating environmental quality and protecting ecosystems and public health. Monitoring of VOCs is important for overall human and environmental health [17]; thus, novel methods and techniques for analysis are being developed [22,23].

Gas chromatography-mass spectrometry (GC-MS) is a key tool for detecting BTEX compounds in environmental samples, such as precipitation, where their concentrations are typically low and embedded in complex matrices [24]. Due to its high sensitivity, selectivity, and capacity for consistent compound identification, GC-MS is the method of choice for trace-level detection of these volatile organic pollutants. BTEX monitoring is essential for evaluating atmospheric hydrocarbon deposition and assessing environmental contamination from vehicle and industrial emissions. Because of its robustness, which enables both qualitative and quantitative evaluations, GC-MS is an essential tool for routine surveillance and research studies on air–water interface pollution [11].

A representative sampling point, selected for its typical Central European city conditions, was chosen for this study. In the selected sampling point, urban influences of traffic and industrial activities converged with seasonal meteorological influences. While a multi-site network could provide wider spatial coverage, the focus was on generating a high-resolution temporal dataset under controlled urban conditions to better understand BTEX wet scavenging processes and pollution dynamics.

The present study aims to assess the role of precipitation in the wet scavenging of BTEX compounds from the urban atmosphere and to explore the meteorological mechanisms that influence their transfer. In addition, it examines how meteorological factors such as wind, temperature, and humidity influence BTEX concentrations in precipitation and evaluates the environmental risks associated with their deposition. Special focus is placed on identifying pollution episodes and potential long-range transport of BTEX through trajectory modeling. By combining chemical analysis, meteorological observations, and air mass back-trajectory simulations, this study offers an integrated framework for understanding BTEX dynamics in urban precipitation. The outcomes contribute not only to characterizing seasonal pollution patterns in a representative Central European city but also provide a foundation for advancing risk-informed environmental monitoring and urban air quality management strategies.

2. Materials and Methods

2.1. Study Area

Novi Sad, the second largest city in Serbia and the capital of the Autonomous Province of Vojvodina, has about 330,000 inhabitants and covers an area of 102 km² on the southern and eastern banks of the Danube, in the northern foothills of Fruška Gora Mountain. The university campus, as one of the key areas with high daily mobility, is rapidly expanding, bringing challenges such as environmental pollution. The proximity of industrial zones, heating plants, landfills, and refineries contributes to the accumulation of BTEX compounds in the urban atmosphere, making the assessment of their concentrations important for identifying potential health threats, the health protection of students and employees, and ensuring a safe and healthy campus environment.

The systematic sampling of precipitation started on 5 September 2024, and continued through 31 December 2024, on the terrace of the BioSense Institute (45.24518° N, 19.85444° E) at an elevation of 20 m above the topographic surface, as explained in Figure 1. The BioSense Institute is located on the campus of the University of Novi Sad in the southeastern part of the city. The institute is located near the bank of the Danube River, adjacent to a pedestrian zone along the river embankment to the east. To the north, it is bordered by a park with poplar trees averaging around 30 m in height, nearby low-traffic roads, and densely packed parking areas. To the south lies a recreational area, while to the west are surrounding faculty buildings, ranging from approximately 3 to 9 stories in height (Figure 1a,b).

2.2. Sampling

The collection of precipitation was conducted employing passive collectors—i.e., high-purity glass containers and a glass funnel with an opening diameter of 24.5 cm—to ensure contamination-free sampling. This approach is based on the study by Grynkiewicz et al. [25], which focuses on atmospheric precipitation sampling techniques and emphasizes the importance of using inert materials to prevent sample contamination. The study recommended using collection devices such as glass bottles and funnels, highlighting their suitability for collecting precipitation samples intended for trace-level analysis.

Precipitation collection was carried out daily for 24 h, starting at 10 a.m. In total, 23 samples were collected in the period from 5 September to 26 December 2024, covering autumn and winter months, distributed as follows: September (n = 5), October (n = 5), November (n = 4), and December (n = 9). The sampling frequency depended entirely on the occurrence of precipitation events, resulting in more samples during months with more frequent or intense rainfall. In order to ensure consistency and comparability of results, for analytical purposes, a standardized volume of 100 mL was extracted from each sample. Although the sampling did not include spring and summer months, the collected data provide valuable insights into seasonal patterns during colder periods and lay the foundation for future year-round monitoring.

2.3. Meteorological Data

To complement the sampling setup, an automatic meteorological station, Mi-Sol model WS-WH24C, was strategically positioned at the site to continuously monitor atmospheric conditions (air temperature, relative humidity, wind speed, wind gust, precipitation) throughout the study period. To supplement this dataset, a meteorological dataset for the period of observation was collected from the Novi Sad Rimski Šancevi official weather station (19°51 E, 45°20 N), located 7.8 km to the north of the BioSense Institute. The dataset from the Rimski Sancevi weather station contains information on average, minimum, and maximum daily values of air temperature, relative humidity, precipitation, atmospheric pressure, wind speed, and wind gusts. Access to measured meteorological data is enabled through the AgroSense platform [26].

2.4. Analytical Methods

BTEX mixture standards, including benzene, toluene, xylene, and ethylbenzene, were supplied by Sigma-Aldrich (St. Louis, MO, USA). All standard solutions were prepared by diluting a stock solution with HPLC-grade n-Hexane. All standards were prepared by a diluted working standard solution with HPLC-grade n-Hexane.

In the present study, precipitation samples, collected in glass bottles, were analyzed by the gas chromatography-mass spectrometry (GC-MS) analytical method. For extraction, a liquid–liquid method was used as described by Paz Otero et al. [27]. Treated precipitation samples were stored in glass bottles at +4 °C until analysis. A 100 mL precipitation sample from the bottle was transferred to a separating glass funnel and extracted with dichloromethane (2 × 30 mL). The respective dichloromethane (DCM) extracts were pooled and evaporated with a vacuum evaporator. The resulting extracts were dissolved in 500 µL of hexane for GC-MS analysis [27]. The experimental equipment included a gas chromatography-mass spectrometry Agilent 7890A/7694E with a fused silica column (30 m × 0.25 µm film thickness HP-5M) from Agilent Technologies Inc. (Santa Clara, CA, USA).

BTEX compounds were identified by matching both retention times and characteristic mass fragments (target and qualifier ions) with those of an external certified standard mixture. Identification was confirmed by comparing the acquired spectra with NIST Mass Spectral Library data and required a match factor above 85% for positive compound confirmation. Quantification was carried out using an external calibration curve using standard solutions of the analyzed compounds. Blank samples were also subjected to analysis and quantified, and GC-MS analysis was performed in splitless mode with a constant flow. Helium was used as carrier gas (v = 35.698 cm/s; p = 7.0 psi).

The instrument conditions for BTEX analysis were injection temperature of 280 °C, MSD 280 °C; and oven with an initial temperature of 40 °C (held for 3 min), ramping up to 150 °C at 20 °C/min (held for 1.5 min).

To ensure the correct identification and accurate quantification of the target compounds, strict quality control was conducted in this study. Preparation of glassware was conducted by washing with deionized water followed with heating in a muffle oven at 400 °C for over 4 h to reduce potential contamination.

All quantitative analyses were performed using a validated in-house method based on external calibration curves prepared from certified standard solutions for each BTEX compound. No internal standard was used in this study. The calibration curve was in the concentration range of 5.0 µg/L to 500 µg/L (5 calibration points). The calibration curves for all compounds exhibited a correlation coefficient (R2) of ≥0.9924. The limits of detection (LOD) and limits of quantification (LOQ) in µg/L were calculated by tripling the signal-to-noise ratio (S/N), as explained in Table S2.

2.5. Data Processing and Statistical Analysis

In order to process the collected dataset, several statistical programs were used. QtiPlot (version 0.9.8.9) [28] and Excel 2010 [29] were used for plotting and statistical analysis. The Spearman’s correlation coefficient heat map was created using the Python programming language 3.13.2 [30] with the pandas, matplotlib, seaborn, and numpy libraries. For the visualization of spatial data, QGIS 3.40.4 was used [31]. Backward trajectories of air masses were computed using the HYSPLIT-WEB model.

2.6. Health Risk Assessment

The environmental risk assessment of BTEX was carried out with an emphasis on aquatic organisms, because rainwater contaminants that end up in aquatic ecosystems mainly represent a threat to aquatic organisms [1,12].

Environmental exposure is usually calculated via the risk quotient (RQ), which is based on a comparison of the predicted no-effect concentration (PNEC) and the predicted environmental concentration (PEC) [1]:

RQ = PEC/PNEC,

(1)

where PNEC is calculated as the ratio of the toxic effect level (TEL) to the assessment factor (AF), i.e., PNEC = TEL/AF. TEL evaluation is based on a choice between LC50 (median lethal concentration), EC50 (median effect concentration), and NOEC (no-effect concentration), i.e., TEL = {LC50 or EC50 or NOEC}.

Here, the actual measured concentration of the analytes in the sample was taken for the PEC values, and TEL data were obtained from OGP (2002).

Table 1. Risk quantification inputs for BTEX in aquatic environments.

BTEX Compound	Endpoint	Trophic Level	TEL (µg/L)	AF	PNEC (µg/L)
Benzene	NOEC (20 d)	Crustaceans	170	10	17
Toluene	LC50 (96 h)	Crustaceans	490	1000	0.49
Ethylbenzene	NOEC (21 d)	Crustaceans	1000	10	100
p-m-Xylene	LC50 (96 h)	Fish	1200	1000	1.2
o-Xylene	LC50 (96 h)	Fish	1200	1000	1.2

Typical risk classification by RQ value is as follows: very high risk (RQ ≥ 10), high risk (10 ˂ RQ ≥ 10), medium risk (1 ˂ RQ ≥ 0.1), low risk (0.1 ˂ RQ ≥ 0.01), and negligible risk (RQ ≤ 0.001) [1,32].

provides all input information necessary to calculate RQ with Equation (1).

The environmental effect assessment of an aquatic system requires test results using entities representing three trophic levels: algae, Daphnia, and fish [33,34]. According to the EC Technical Guidance Document of the European Commission [33], an assessment factor of 1000 is used for the determination of PNEC from the three trophic levels of algae, Daphnia, and fish in case the data from the acute testing are the only available data. AF can be refined to 100, 50, or 10 if the values of NOECs are available from long-term tests covering one, two, or three trophic levels.

3. Results and Discussion

3.1. BTEX in Precipitation

A set of five BTEX concentration time series obtained from 23 precipitation samples during the period of observation between 5 September and 25 December of 2024 is presented in Figure 2, along with corresponding risk quotient time series. The RQ time series were generated with Equation (1) by using the inputs from Table 1.

Benzene, ethylbenzene, p-xylene, and o-xylene showed a consistent two-phase pattern: concentrations remained stable from the beginning of the observation until 22 November, then dropped to below the method detection limit (0.1 µg/L) after 28 November. The pink rectangular areas on the plots indicate the period between the phases. Prior to the drop, mean concentrations were 0.45, 3.45, 4.09, and 2.31 µg/L, respectively, with minimal temporal variability (coefficients of variation, CV: 0.9%, 2.3%, 2.4%, and 4.2%). After 28 November, values were indistinguishable from zero. This transition suggests the influence of either a sharp change in emission intensity, a meteorological shift, or a combination of both, which is further explored in the subsequent sections. The risks were classified as high for both xylene isomers before the drop, while benzene and ethylbenzene remained low risk throughout the observation period.

The toluene time series deviated from this pattern. Its concentration was stable before 22 November (mean 1.32 µg/L, CV 2.3%), then increased sharply to exceed 6 µg/L. Between then and 9 December, it gradually declined to a new quasi-stable level of 1.12 µg/L, which was still above the detection limit. Such dynamics can be explained by emission from regional sources in combination with meteorological conditions such as a northwest wind, which can be responsible for the toluene transfer during the period of elevated concentration, as further investigated in Section 3.3. The corresponding risk level was high throughout the study, with an exceptional peak on 28 November, where the risk was classified as very high.

Table 2. Summary of BTEX concentration data with respect to precipitation grouped by time frames with corresponding risk quotient estimations.

Anayte	Mean, µg/L	Stdev, µg/L	RQ, n.u.	Mean, µg/L	Stdev, µg/L	RQ, n.u.	Stat Test
	05/09–22/11			28/11–25/12
Benzene	0.45	0.004	0.03	0.08	0.004	0.005	p < 0.01
Ethylbenzene	3.45	0.08	0.03	0.03	0.005	0.0003	p < 0.01
p-Xylene	4.09	0.10	3.41	0.02	0.005	0.02	p < 0.01
o-Xylene	2.31	0.10	1.93	0.04	0.01	0.03	p < 0.01
				28/11–29/11
Toluene	1.32	0.03	2.70	5.34	----	10.89	Z = 6.3
				09/12–25/12
				1.12	0.19	2.28	p = 0.01

Stdev—Standard deviation.

summarizes the BTEX data categorized by the time frames, where the p-value indicates the result of a one-way ANOVA test for equality of means and the Z-score is calculated based on the toluene concentration mean and standard deviation for the period before 22 November.

The BTEX concentration levels presented here agree with several studies [3,11,35] showing high VOC emissions in urban areas. These findings suggest that toluene and xylene isomers are the most prevalent BTEX compounds at the current location and season in terms of the environmental risks associated with their presence in precipitation. The elevated levels of p-xylene and o-xylene can be explained by their wide use in industrial solvents, fuel mixtures, and comparatively lower reactivity with atmospheric hydroxyl radicals compared to benzene and toluene [36]. The variation in xylene concentration in precipitation can be caused by temporal and/or spatial heterogeneity in emission sources such as traffic surges and industrial production, along with meteorological influences [37]. Given their persistence in the air and strong affinity for wet deposition, these compounds may serve as reliable indicators of episodic urban pollution under specific meteorological conditions.

Toluene emission sources are the production, transport, and use of gasoline. Motor vehicle exhaust and evaporative losses from fuel storage and service stations are major contributors to air toluene levels. Additionally, industrial processes and the use of toluene-based products like paints and adhesives also contribute to emissions. The highly elevated toluene concentration measured on 28 November may reflect episodic or point-source pollution incidents. Such episodic maximum events have also been reported correspondingly in city air observations in temperature inversions or when wind is stagnant [38].

Both benzene and ethylbenzene are of less concern in terms of the risks demonstrated in the present study. Ethylbenzene usually serves as a stable urban pollution marker, which is less volatilized compared to toluene and stays for long enough in the atmosphere to get scavenged by rain [39]. Benzene, which had the lowest concentration level in all samples, is often reported as a dominant VOC in urban atmospheres due to its ubiquity in fuel and combustion processes. As a relatively well-soluble compound in water (Henry’s law constant), it is less prevalent in precipitation per this research, likely because of reduced emissions within the region or increased atmospheric degradation rates. These findings are consistent with previous observations in urban China and Europe, where control strategies have reduced benzene emissions in the past decade [38,40].

The observed differences in concentration profiles also align with known physicochemical properties of BTEX compounds. For instance, xylene isomers generally have lower water solubility and higher vapor pressure than benzene, which could influence their atmospheric behavior and wet deposition efficiency [41]. Additionally, meteorological conditions such as rainfall intensity, temperature, and wind direction may further affect the scavenging and dilution of BTEX compounds in precipitation [42]. Through previous studies, climatic variables such as temperature, humidity, wind speed, and precipitation rate have been proven to significantly influence the dispersion and removal of BTEX compounds within urban atmospheres. Temperature and humidity are key parameters influencing the volatilization of BTEX compounds and air diffusion. At elevated temperatures, the volatility of such chemicals is higher, leading to atmospheric concentrations that are more likely to be entrained by rain. Lower temperatures, however, may suppress their volatility, modifying their distribution in the atmosphere and precipitation. Moreover, wind speed has a significant role to play in spreading BTEX compounds. Greater wind speeds carry VOCs longer distances, producing broader spreading of contamination, while smaller wind speeds permit VOCs to be carried near their point of origin. Precipitation is important as well; it scavenges suspended VOCs from the air and deposits them into the ground via raindrops, thereby eliminating them. The observed pattern of BTEX concentration time series with a distinct transition between two phases suggests the importance of meteorological triggers in rapidly altering the atmospheric exposure of BTEX compounds. In particular, pressure gradients, wind gusts, and precipitation likely act synergistically in BTEX removal.

Considering all these factors, the data presented here might be explained by changes in urban or industrial emission levels, weather conditions, or a combination of both. In the absence of reliable urban pollution data, this study examined the potential meteorological influence on BTEX concentrations by analyzing weather data collected during the observation period. More specifically, the following three questions were addressed here: (1) What caused the removal of benzene, ethylbenzene, and xylene isomers from the air after 28 November; (2) why did the concentrations of these compounds drop quickly rather than decrease gradually; and (3) what caused the short-term elevation of toluene.

3.2. Meteorological Impacts on BTEX Levels

According to [43], the study area belongs to the open midrise local climate zone (LCZ). In brief, the average autumn and winter air temperatures are 1.4 °C higher compared to natural areas outside the city, which indicates the presence of an urban heat island effect [44]. The area is characterized by high humidity, which could be linked to its morphological structure (a mix of impervious surfaces such as buildings and boulevards with the green zones) and prolonged moisture retention compared to natural areas outside the city [45]. According to the Novi Sad municipality air quality report, the highest levels of air pollution in the urban environment are observed during the winter months, where the main sources of pollution are intensive use of fossil fuels for heating (coal, oil, and gas) and increased emissions from traffic. Winter meteorological conditions are characterized by frequent temperature inversions and low winds, which prevent the dispersion of pollutants and allow their accumulation in the lower layers of the atmosphere. As a result, the pollutants do not reach the troposphere, which may lead to reduced concentrations in precipitation during the colder seasons. The interaction between the urban heat island effect and temperature inversions during late autumn could contribute to pollutant entrapment at lower atmospheric levels. This stratification may reduce vertical mixing and limit BTEX transport into the troposphere, hence altering the efficiency of wet deposition processes [46].

The meteorological data collected for the period of observation and selected for the analysis included daily average values of air temperature, relative humidity, atmospheric pressure, precipitation, and wind speed, along with daily minimum air temperature, wind gusts, and directions. The corresponding time series with box-and-whisker plots are presented in Figure 3 and Figure 4. Here, red vertical dash lines border the period between two phases of BTEX concentration behavior in the time series plots. All box-and-whisker plots are meant to compare four-week data samples that correspond to the time frames before 22 November and after 28 November, where the p-value in the corner indicates the result of Welch’s t-test for the equality of sample means. Eight data points that belong to the period between 22 November and 28 November are also shown in between the boxes as red empty circles labeled “Drop” to highlight the fact that four of five BTEX compounds are dropped during this period.

Air temperature followed a decreasing trend during the period of observation. Unsurprisingly, it negatively correlated with humidity (Pearson’s coefficient of −0.67), which followed an increasing trend. Both temperature and humidity values were significantly different for the time frames before and after the drop. The daily minimum temperature reached zero on 4 November, whereas the daily average temperature reached zero on 16 November. Precipitation was more frequent in the period after 28 November, but with lower intensity. A local high precipitation value of 33 mm must be noted as a remarkable precipitation event on 23 November, i.e., within the “Drop” time frame. Given the intensity and timing of this rain event, it is reasonable that a rapid atmospheric scavenging of BTEX compounds occurred, especially if preceded by accumulation under stagnant conditions. Episodic rains following periods of calm can act as efficient scrubbers for semi-volatile organic compounds such as BTEX [11,39].

Atmospheric pressure was not statistically different for the periods after 28 November and before 22 November. However, the pressure time series curve has a specific pattern near the “Drop” time frame. Indeed, atmospheric pressure gradually decreased for 7 consecutive days, from 13 November to 20 November, when it reached the absolute minimum from the start of observation. As the “Drop” period began, the pressure increased rapidly from a minimum value of 988 mbar on 22 November to 1019 mbar on 24 November, thus indicating the highest change detected for the entire period of observation.

Based on standard meteorological classification of wind strength (the Beaufort scale) [47], winds classified as moderate breeze (average speeds of 20–29 km/h) can lift dust from the ground surface and bring VOCs from longer distances. Such winds were recorded only three times during the observation period (18 October, 20 November, and 22 November), with the strongest occurring on 20 November and 22 November. (The period from the very beginning of November until the occurrence of moderate winds was characterized by lighter wind conditions, and probably had more local character.) The rest of the registered winds can be classified as follows: light wind (12–19 km/h) with 24 occurrences, light breeze (6–11 km/h) with 65 occurrences, and light air (1–5 km/h) with 14 occurrences.

The daily average wind speed was statistically higher during the period after 28 November than during the period before 22 November. The highest daily average wind speed (>30 km/h) was detected on 22 November. This wind was responsible for the rapid atmospheric pressure increase followed by rain. Three strong wind gusts were detected between 22 November and 28 November. Notably, all three gusts were from the northwest direction. The rapid increase in atmospheric pressure and the onset of strong northwestern winds may have triggered large-scale air mass exchange, replacing stagnant polluted air with cleaner inflow, thus resetting BTEX concentrations in the local atmosphere [48,49].

In summary, the periods before 22 November and after 28 November differ significantly in terms of air temperature, humidity, and rain events, which could explain the change in BTEX compound concentrations in precipitation. Indeed, increased air temperatures usually contribute to more intensive evaporation of BTEX compounds in the air, while temperature inversions, which are common during the cold season, can trap pollutants near the ground, not letting them be transported to the troposphere. According to the Weather Spark portal, mists and fog, which are probable indicators of the inversions, were observed for 15 and 12 separate dates in November 2024 in Novi Sad, respectively. Humidity usually increases BTEX adsorption on solid particles in the atmosphere, which can reduce the concentration in the gas phase of the air. Precipitation scavenges BTEX compounds from the air since aromatic hydrocarbon vapors adsorb to the air/water interface and are transported by wet deposition processes via rain [39,50].

The time frame between 22 November and 28 November can be characterized by a unique meteorological pattern when compared to the period before. This may help to address the question related to the dramatic change in benzene, ethylbenzene, and xylene isomer concentrations in precipitation. Each individual notable event which happened during this period (the fastest pressure change on 22 November, the strongest daily wind the same day, the heavy rain on 23 November, and the series of three consecutive strong northwest wind gusts) or a combination of them could be responsible for triggering the shift between the two BTEX-related weather conditions. For example, the 23 November rain, when combined with the intense air mass transport, could have scavenged all four compounds from the troposphere, while their regular transfer back to the air from the urban emission sources was blocked by the low temperatures and high humidity. Alternatively, a spatial shift in dominant wind direction could have imported air masses with different pollutant signatures. If regional industrial zones located northwest of the city released higher amounts of toluene during this period, wind-assisted transport could explain the sudden peak [51,52]. As for the elevation of toluene on 28–29 November, it is possible that additional toluene reached the Novi Sad area from another location, potentially carried by the northwest winds. To elaborate on such a hypothesis, air mass flow mathematical modeling was applied (see Section 3.3).

3.3. Modeling Air Mass Transport Using the HYSPLIT Model

Analysis of the measured BTEX concentrations revealed that only toluene exhibited elevated levels that could pose very high environmental risks and threaten the health of aquatic organisms. The highest toluene concentration was recorded on 28 November, associated with very high risks. Accordingly, backward trajectory modeling was conducted for the period of 28–29 November 2024.

In order to identify a potential source zone for toluene, the online version of one of the most widely used models for atmospheric trajectory and dispersion calculations of pollution, the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT), was applied [53,54]. The modeling involved the selection of backward trajectories and the use of the Global Data Assimilation System (GDAS), which encompasses the historical period from 2006 to the present, with 1 degree resolution.

In the case of the analyzed period from 28 to 29 November, the total run time (hours) was set to 24 h, and the trajectory generation time was set to every 6 h. The model predicts the generation of a total of four trajectories. Considering the recommended heights used for air pollution modeling using the HYSPLIT model, each of the four trajectories is set to start its path at the appropriate height within the defined three levels of height: 100 m, 500 m, and 1000 m AGL. At 500 m, AGL is usually the height of the atmospheric boundary layer [9], while at 100 m, AGL is the most intense impact of surface pollution [55,56]. The height of 1000 m AGL is close to the upper planetary boundary layer and transitions to the free troposphere, where all pollution can be transported over long distances [57,58]. The results of the model are shown in Figure 5.

By examining the time of recorded precipitation using the Mi-Sol weather station, it can be concluded that there was no registered precipitation between midnight and 4 p.m. on 28 November. During this period, the Novi Sad area was under the dominant influence of air masses observed in the lower atmosphere (represented by the light blue and green trajectories). Starting from 5 p.m., the station recorded the first precipitation. At that time, according to Figure 5, air masses from three trajectories—green, light blue, and dark blue (the latter descending from altitudes above 1500 m AGL)—simultaneously reached the Novi Sad area. The highest amount of precipitation was recorded between 6 p.m. and 7 p.m. Another period of increased precipitation was observed around 3 a.m. the next day, 29 November. According to Figure 5, at that time, Novi Sad was under the influence of air masses associated with the red trajectory. Based on the obtained trajectory model, the potential source of pollution associated with the blue and red trajectories is the area over Central Europe. In this region, the air masses converge and concentrate over northeastern Croatia and southwestern Hungary, where the green and light blue trajectories also appear to originate. Most trajectories approached the Novi Sad area from the northwest, with the exception of the green trajectory, which came from the west.

Looking at the hourly distribution of wind directions and peak gusts (Supplementary information, Figure S2), the strongest wind gusts on 28 November 2024, were recorded at 3 p.m. and 4 p.m. The direction of the gusts was northwest–southeast, and based on their timing, the descending flow of air masses along the dark blue, light blue, and green trajectories would correspond to this event. Since precipitation started around 5 p.m., it is very likely that these rain events were responsible for washing out pollutants from the atmosphere. The second strong gust was recorded in the early morning hours of 29 November (around 8 a.m. and 9 a.m.), along with the registered precipitation. This event corresponds to the red trajectory from Figure 5. In summary, there is no single air mass that can be identified as the only source of toluene pollution. Rather, it is likely a combined contribution, as all trajectories predominantly originate from the northwest direction and arrive in relatively close time intervals in Novi Sad, Serbia.

Given the clear shift in wind direction and the convergence of multiple air masses from NW and WNW directions, it is conceivable that the spike in toluene concentration was driven by an episodic emission event outside the immediate urban area. Such an event could include accidental release, unreported industrial flaring, or intense use of solvent-based products under stagnant air conditions. Moreover, the descending trajectory from >1500 m AGL on 28 November raises the possibility that aged air masses, enriched with VOCs from distant sources, reached Novi Sad through advection. This hypothesis aligns with observations in similar European cities where sudden spikes of toluene have been linked to cross-border transport during stagnant weather episodes.

Given the complex pattern of air mass trajectories converging on the Novi Sad area, it is plausible that multiple sources contributed to the elevated toluene levels, rather than a single dominant emitter. The episodic nature of the concentration spike suggests that transient events, such as accidental industrial releases or undocumented flaring activities, might have played a critical role. Such events are notoriously difficult to capture in routine monitoring, especially under stagnant meteorological conditions that inhibit pollutant dispersion and facilitate accumulation near the surface [57].

Moreover, the influence of aged air masses descending from altitudes above 1500 m AGL implies that long-range atmospheric transport cannot be discounted. This advective transport may introduce volatile organic compounds (VOCs), including toluene, originating from distant industrial or urban centers across Central Europe. This mechanism aligns with documented cases in other European urban environments where episodic VOC peaks were traced back to cross-border pollution episodes exacerbated by stagnant atmospheric conditions and temperature inversions [58,59].

The coincidence of precipitation events shortly after the wind direction shift further supports the hypothesis that wet deposition may have contributed to the sudden reduction in ambient toluene concentrations, indicating a potential episodic washout of accumulated pollutants. However, the timing also raises questions about the interaction between meteorological phenomena and pollution dynamics, warranting more detailed studies using higher temporal resolution data and combined chemical transport models [60].

Finally, these findings underscore the challenges in source apportionment of atmospheric pollutants in regions influenced by complex meteorology and multiple emission sources. They highlight the importance of integrating trajectory modeling with comprehensive emission inventories and real-time industrial activity data to better capture the episodic nature of pollution events and improve regional air quality management strategies [61] (Table S1).

3.4. Future Perspectives and Limitations

The present work is a pilot study that aims to develop and validate analytical methodology with respect to rainwater collected at a single sampling location. Future research should expand to a multi-site design that captures spatial variability across the region. To distinguish between VOCs detected locally and those transported over long distances, it will be crucial to employ advanced techniques such as isotopic tracing or source apportionment modeling. Additionally, year-round monitoring across multiple seasons will provide a deeper understanding of BTEX dynamics under fluctuating atmospheric conditions.

4. Conclusions

This study provides a detailed assessment of the presence and behavior of BTEX compounds—benzene, toluene, ethylbenzene, and xylene isomers—in urban precipitation, based on samples collected in Novi Sad during autumn and winter 2024 and analyzed using gas chromatography-mass spectrometry. By integrating chemical analysis with meteorological observations and backward trajectory modeling, the research proposes the mechanisms driving BTEX removal from the atmosphere via wet scavenging and highlights the influence of both local weather patterns and long-range atmospheric transport on pollutant concentrations.

Higher BTEX levels were recorded during the early observation period (September to late November 2024), with concentrations of benzene, ethylbenzene, p-xylene, o-xylene, and toluene averaging 0.45, 3.45, 4.0, 2.31, and 1.32 µg/L, respectively. These values sharply declined to near-zero levels by December 2024 for all compounds except toluene, which remained at 1.12 µg/L. A significant finding was the abrupt and simultaneous drop in benzene, ethylbenzene, and xylene concentrations in late November, coinciding with notable meteorological changes—rapid increases in atmospheric pressure, strong northwesterly winds, and intense rainfall. Conversely, toluene exhibited a pronounced spike above 6 µg/L on 28 November, likely resulting from transboundary transport, as confirmed by HYSPLIT trajectory simulations tracing air masses back to Central European regions. These observations underscore precipitation’s dual role as both a scavenger of atmospheric pollutants and a conduit for their redistribution into local ecosystems.

Overall, the results indicate precipitation as a great integrative matrix for tracking atmospheric VOCs, particularly in urban settings. It captures short- and long-term patterns of pollution, reacts to emission source variability, and offers complementary approaches to standard air monitoring. The findings support the need for continuous environmental monitoring, considering that BTEX emissions are changing and have implications for air quality as well as public health. Taken together, the results underscore the importance of considering both compound-specific properties and environmental dynamics when evaluating BTEX pollution patterns in precipitation.

BTEX compounds pose a serious environmental risk, particularly due to the high risk levels identified for xylene isomers and toluene in this study. These findings indicate the need for a broader regulatory approach that goes beyond the control of air concentrations alone. While current regulations are primarily focused on air quality and industrial emissions, it is evident that precipitation should also be considered as an important pathway for the introduction of pollutants into the environment.

In conclusion, effective environmental protection requires accounting for the complete transport and deposition cycle of hazardous pollutants. Incorporating monitoring and control of BTEX compounds in precipitation into existing policies will foster a more holistic approach to protecting air, water, and ecosystem health across Europe.

While current data do not conclusively distinguish between local vs. regional emission sources, the patterns observed underscore the necessity for high temporal-resolution monitoring and real-time source attribution systems. A single-day pollution episode, as observed with toluene on 28 November, can pose acute ecological risks, particularly when occurring in coincidence with rainfall. Future studies should aim to include real-time industrial activity logs, traffic data, and vertical air profile measurements to disentangle the relative contribution of emissions versus meteorological phenomena, and to explore the use of satellite-based emission detection and predictive air mass modeling to anticipate such events in advance and inform early warning systems.