Substantial efforts and improvements in air quality across Europe lowered levels of air pollutants including particulate matter (PM) in the last decades. However, significant proportion of the European population still lives in areas exceeding WHO recommendations, especially in Northern Italy, Balkans, and Eastern Europe, including Serbia. Targeted PM mitigation strategies require extensive air quality monitoring and modelling including source apportionment (SA) studies. In the past, numerous SA studies were conducted for Belgrade city, capital of Serbia. Nevertheless, comparisons across the results are difficult, as they encompass different datasets of pollutants contained in PM fractions such as elements and/or ions and/or PAHs. Here, the aim is to offer a broader insight on PM10 sources at an urban background site in Belgrade by including 34 species as input variables for SA (carbonaceous aerosols, elements, ions and specific organic tracers). For SA, the USEPA PMF 5.0 software was applied. The factor that dominated PM10 mass was biomass burning (21 %), primarily during heating season, followed by ammonium sulphate (18 %) and mineral dust (17 %). A mixed traffic and industrial activity accounted for 15 % of PM10 mass while contribution of factors from biological origin, primary biological aerosol particles and biogenic secondary organic aerosols from isoprene, was 10 % each. Mixed factor of long-range transport and road salt/local combustion contributed to the PM10 mass 9 %. This analysis provides more detailed perspective on the composition and sources of PM10 in Belgrade, both from anthropogenic and natural, including biological origin. These findings are valuable for defining targeted PM10 mitigation strategies.
European Commission101060170Ministarstvo Prosvete, Nauke i Tehnološkog Razvoja451-03-33/2026-03/ 200017Introduction
The World Health Organization (WHO) estimated that 4.2 million premature deaths at the global level are caused by exposure to outdoor air pollution (World Health Organization, 2024). Among the air pollutants, particulate matter (PM) represents the biggest health concern. PM may originate from both anthropogenic and natural sources. These sources emit particles that vary in size, morphology, and chemical composition, and these differences influence how particles may be harmful for human health (Liu et al., 2025). Coarse particles (<10 and >2.5µm) primarily deposit in the extra thoracic airways, whereas fine (<2.5µm) and ultrafine particles (10–100 nm) (Ruuskanen et al., 2001) can penetrate deeper into the pulmonary region of respiratory tract (Khan et al., 2022; Zhai et al., 2024). Exposure to particles lower than 10 µm (PM10) is related to increased mortality and morbidity from cardiovascular and respiratory diseases (Polichetti et al., 2009; Newell et al., 2017). In 2016, the International Agency for Research on Cancer (IARC) classified outdoor air pollution, as well as PM, as carcinogenic to humans (group 1) based on sufficient evidence in both humans and experimental animals. According to IARC, exposure to PM contributes to the development of lung cancer (IARC, 2016).
Although substantial efforts and improvements in air quality across EU significantly lowered levels of air pollutants (especially SO2, NOx) in the period from 2000 to 2019 (Aas et al., 2024), there is still important proportion of the European population that lives in areas exceeding WHO recommendations (European Environment Agency, 2025a). The most severe situation is in low- and middle- income countries where is estimated that citizens are exposed to up to four times higher levels of ambient PM2.5 (Health Effects Institute, 2024). Cohen et al. (2017) demonstrated that even modest increases in PM concentrations can result in disproportionately large increases in health risks, underscoring the need for further mitigation. Moreover, substantially stronger evidence now indicates that air pollution may impact human health at lower exposure levels than previously recognized (Stafoggia et al., 2022). Therefore, the WHO established new guidance levels in 2021 where new set of annual and daily average levels of PM10 and PM2.5 are lowered and may be used by policy makers who are responsible for air pollution control. WHO recommended average annual PM10 and PM2.5 levels are 15 and 5 µgm-3 while average 24 h target levels are 45 and 15 µgm-3, respectively. Many studies demonstrated correlation between exposure to PM2.5 and cardiopulmonary and lung cancer mortality, making it the world's leading environmental risk factor (McDuffie et al., 2021; Burnett et al., 2018; Pope et al., 2002). However, some authors highlight that the health effects from coarse particulate matter should not be overlooked as the adverse health effects related to this fraction distinct beyond the attributable effects of PM2.5 alone. Therefore, the continuous monitoring of PM10 is necessary as it includes both coarse and fine fraction (Tian et al., 2020; Choi et al., 2026).
Targeted airborne PM mitigation strategies (<10µm) require extensive air quality monitoring and modelling including source apportionment studies. Therefore, new EU Air Quality Directive underlines the importance of monitoring at supersites novel pollutants such ultrafine particles, black carbon and elemental carbon, ammonia and oxidative potential (European Council, 2024). Directive emphasize that it is crucial to systematically monitor air quality in areas where pollutants concentrations are high, and measurements are scarce. Annual PM10 levels in Serbia in 2023 and 2024 exceeded EU limit value (40 µgm-3) on several monitoring sites in Belgrade, Valjevo, Užice, Zaječar, Novi Pazar and Šabac. In 2023 and 2024, PM10 was the pollutant with the highest number of exceedance days in the Republic of Serbia (Serbian Environmental Protection Agency, 2024, 2026). European Environment Agency reports show that Serbian PM10 levels in 2023 were higher than European averages (European Environment Agency, 2025b) and Health Effects Institute report that air quality monitoring data is still scarce in this area (Health Effects Institute, 2022). Liu et al. (2025) emphasised that sources like secondary aerosols and road traffic, show consistent chemical profile, temporal variability, and contribution to PM10 concentration in source apportionment studies conducted in twenty-four cities across Europe. However, emissions from industry, heavy oil combustion, and crustal sources vary considerably by location, making local source apportionment essential for effective pollution control.
Source apportionment studies using different approaches were also conducted in Belgrade Metropolitan area, collectively covering the sampling period from 2003 to 2015. The sources most commonly identified were coal and wood combustion, secondary aerosols, traffic (vehicle exhaust) and mineral matter (Mijić et al., 2012; Cvetković et al., 2015; Stojić et al., 2016; Todorović et al., 2020). The influence of marine aerosol, construction activities, fertilizer plant, oil combustion and industry were observed too (Rajšić et al., 2008; Đorđević et al., 2012) (Table S1, Fig. S1 in the Supplement). In these studies, source with the biggest contribution to PM concentration was coal and wood combustion which authors mainly attributed to the thermal power plants “Kolubara A” and “Nikola Tesla A and B”, located in the Belgrade Metropolitan area, tens of kilometres from the central zone, as well as the usage of coal and wood for district and household heating. Đorđević et al. (2012) observed that the dominant ionic specie in 2008 in Belgrade was formation of ammonium sulphate. Sulphate in the atmosphere mainly originates from the oxidation of SO2 in urban areas and ammonium ion originates from biogenic NH3 but also from fertilizer plant located in Pančevo, 15 km away from Belgrade (Đorđević et al., 2012). However, the fertilizer plant in Pančevo was out of work during WeBaSOOP campaign. In the period when the majority of the abovementioned studies were conducted, the consumption of natural gas for district heating was only 30 % of the total gas consumption in Serbia (period 2001–2010) (Ivezić et al., 2016). Today, natural gas is the main fuel used for district heating in “Belgrade Power Plants” (https://beoelektrane.co.rs/zastita-zivotne-sredine/, last access: 8 August 2026). Further, by 2020, around 1300 heating boilers within the “Belgrade Power Plants” system had been shut down and reoriented to operate on natural gas (https://beoelektrane.co.rs/o-nama/osnovni-podaci/, last access: 8 August 2026). In 2024, the flue gas desulphurisation plant at “Nikola Tesla A” started operating which resulted in reduction in SO2 emissions (https://www.eps.rs/eng/Pages/Istorija.aspx, last access: 8 August 2026). The international E-70 highway passes through Belgrade but today, it functions primarily for local transport since the new bypass road around Belgrade now serves as the main transit road (Ćirović et al., 2026).
Abovementioned studies used different receptor modelling (Unmix, principal component analysis, positive matrix factorisation (PMF)), making it difficult to compare obtained results. In addition, sampling sites where daily (24 h) filter samples were collected, were in Belgrade city area, in triangle of 20 km2 as it is presented at Fig. S1. For one of SA study, Cvetković et al. (2015), one site was in the medium of triangle and another two sites were out of triangle nearby thermal power plants in Metropolitan area of Belgrade far away for several tens of kilometers from the city center. Only studies of Todorović et al. (2020) and Almeida et al. (2020) overlap with the period of sampling and sampling site as filter samples were collected at the urban background area at Zeleno brdo (Fig. S1). Furthermore, species used in these SA studies were different (Table S1). Some studies were performed only upon analyses of elements (Mijić et al., 2012; Rajšić et al., 2008; Almeida et al., 2020), ions (Ðorđević et al., 2012) or PAHs (Cvetković et al., 2015). Almeida et al. (2020) for SA study used data collected for 16 months and concentrations of elements while Todorović et al. (2020) as input data set used one year of Almeida et al. (2020) data set and performed additional analyses of ions. In light of this, the present study aimed to perform PM10 source apportionment in Belgrade using PMF, a multivariate receptor model capable of resolving PM mass concentration into contributions of various source types (Paatero and Hopke, 2009) by utilizing error estimates of the species from the data matrix and implements strict non-negativity constraints for the factors (Paatero and Tapper, 1994). Input data set of one yearlong campaign consist of more than thirty species contained in PM10 of atmospheric aerosols in Belgrade. That was for the first time that is perform SA upon data set that contain levels of both carbonaceous aerosols (OC, EC) and organic tracers in the Republic of Serbia and even in wider Western Balkan area. The objective of the study is to offer a broader perspective on PM10 sources by incorporating ions and, elements, and for the first time number of tracers for biogenic aerosols as well as tracers for biomass burning together with OC and EC as input variables, and thereby quantify both natural, including biological, and anthropogenic sources at an urban background site in Belgrade.
Material and methodsSampling site
Our sampling campaign was performed at an urban background monitoring site Ada Marina (44.790166° N, 20.4169° E; 71 m above sea level) which is a part of the local air quality network in Belgrade Metropolitan (Fig. 1) where main air pollutants and meteorological parameters (temperature, relative humidity, wind speed, wind direction and atmospheric pressure) have been measured. Site is located at Ada Ciganlija, the largest recreational area in the central area of Belgrade city, along the Sava River. Possible pollution sources include numerous restaurants in the vicinity of the site places around the lake, biomass burning from the nearby residential area and a high traffic arterial road nearby. In addition, E-70 highway is located approximately 250 m away from the sampling site. Other potential sources of aerosol pollution include a natural gas heating plants for district heating, located about 1.5 km to the Northwest and Southeast. Airport “Nikola Tesla” is 10 km in the Northwest. A municipal landfill at Vinča settlement is 10 km to the East. Another possible pollution source is road traffic associated with 1.7 million citizens and the fact that passenger car fleet in Serbia primarily consist of imported, fossil-fuelled second-hand cars (Mijailović et al., 2019). Furthermore, the coal-fired thermal power plant “Nikola Tesla” is located around 25 km to the Southwest while power plant “Kolubara” is distanced 50 km to the South of the Ada Marina site. Outside the city, there are the areas of agricultural activity as well as Smederevo ferreous smelter and “Kostolac” thermal power plant, 30 and 60 km, respectively, to the East.
Aerosol filter samples
Three simultaneous ambient aerosol filter samples were collected from 6 June 2023 to 28 May 2024, each from three co-located and identical low-volume samplers (LVS3, Leckel, 2.3 m3 h−1) equipped with PM10 inlets. Daily filter samples were collected for 24 h. The field campaign lasted for one year, and for PMF analysis seven representative samples for each month were used which yielded in 84 PM10 samples. The samplers were loaded with quartz fibre filters (PALLFLEX Tissuequartz 2500QAT-UP, 47 mm in diameter), which were conditioned at 20±1 °C and 50±1 % RH for 48 h prior to sampling to enable gravimetric determination of PM10 mass. Filters from one sampler were designated for the analysis of OC, EC, organic tracers, and oxidative potential. Filters from a second sampler were conditioned following exposure for gravimetric analysis, then stored in polystyrene Petri slides at -20 °C until elemental and ionic analyses were performed. PM10 filter samples from a third sampler were retained as backups. All filters were stored at 4 °C prior to exposure and at -20 °C afterward. Field blank filters were collected to assess potential contamination during handling and transport. OC, EC, ions, elements, and organic tracers were analysed from 84 samples.
Organic carbon (OC) and elemental carbon (EC) were measured by thermal-optical analysis using a Sunset Laboratory OC / EC aerosol analyser. The instrument operated according to the EUSAAR-2 temperature protocol (Cavalli et al., 2010) using optical transmission correction for charring.
Measurement of the elements and ions
The elements (Na, Mg, Al, K, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Cd, Ba, Pb, P, Ti, Zr, Mo, Sb, Sn) were quantified utilizing inductively coupled plasma – mass spectrometry (Agilent model 7700), following the procedure described in (Querol et al., 2001). Ions (Cl−, NO3-, SO42-, NH4+) were analysed from the solid phase of suspended particles in the air using ion chromatography (IC). The analysis was performed on Thermo Scientific Dionex™ ICS-1600 ion chromatograph with suppressed conductometric detection.
Measurements of organic tracers
Organic tracers (Table 1) were analysed using ultra-high-performance liquid chromatography (UHPLC; Vanquish UHPLC, Thermo Fisher Scientific) coupled with a Q Exactive™ Plus Orbitrap mass spectrometer (Thermo Fisher Scientific), following the method described by Yttri et al. (2024).
Organic tracers analysed in this study and their source categories.
Organic tracersSourcesLevoglucosan, mannosan galactosanBiomass burningArabitol, mannitol, fructose, glucose, trehalosePrimary biological aerosol particles2-methylerythritol, 2-methylthreitolBiogenic secondary organic aerosol from oxidation of isoprenePositive matrix factorization model
In the present study, the EPA PMF v.5.0 has been used. Our data matrix dimensions were 84×44 (sample number, number of species, respectively). As the input data we used concentration of species obtained from filter analysis and uncertainties calculated following procedure in Norris et al. (2014). Uncertainties were calculated using Eq. (1) if the concentration of the species was higher than limit of detection (LOD) value:
u=(error fraction×concentration)2+(0.5×LOD)2
In the model, if the concentrations of the species were assigned as “below LOD”, a value equal to 1/2 od the LOD was entered. For those species below LOD, the uncertainties were calculated as 5/6 of the LOD (Polissar et al., 1998). Error fraction for all species was 20 %. In this study, signal-to-noise ratio (S/N) and percentage of the samples below LOD (50 %) were used to classify species as “strong”, “weak” and “bad”. Species selected for PMF are reported in Fig. S1. The species that had S/N<1 and samples below LOD >50 % were considered as “bad” (in our study Se, Ga, Co, Zr), while species that had S/N<1 and samples below LOD <50 % and vice versa (S/N>1, below LOD >50 %) were considered as “weak” (Cl−, Cu, As, Cd, Ti, Sn). Species with S/N ratio >1 and samples below LOD <50 % were considered “strong” (OC, EC, NO3-, SO42-, NH4+, Na, Mg, Al, K, Ca, V, Mn, Fe, Rb, Sr, Pb, Sb, galactosan, mannosan, levoglucosan, 2-methylerythritol, 2-methylthreitol, glucose, fructose, mannitol, arabitol, trehalose). Few species, although having a satisfactory S/N and percentage of above LOD, were excluded from the model due to the contamination of the blanks (Cr, Ni, Mo, Ba) or higher analytical uncertainty of the results (Inositol, Erythritol, Zn, P).
To determine the optimal number of factors, solutions with a total number of factors from 3 to 9 with random seed number and no applied factor constrains were examined. The final solution was derived from 100 base runs, with robustness evaluated according to the recommendations of the European guide on air pollution source apportionment with receptor models (Belis et al., 2019a): (1) all factors' chemical profile and temporal variability could have been interpreted, (2) Qtrue/Qexp ratio, (3) normal distribution of modelled species with more than 95 % uncertainty-scaled residuals was within ±3, and (4) on displacement (DISP) and bootstrap (BS) analysis which confirmed the stability of the model.
The input data included sets of chemical tracers that allowed identification of chemical profile and temporal variability connected to certain sources based on previously conducted studies. Therefore, levoglucosan, mannosan, galactosan, K, Rb, OC and EC were tracers for biomass burning (Yttri et al., 2011; Weber et al., 2019), sugars and sugar alcohols for primary biological aerosol particles (Bozzetti et al., 2016) and 2-methyl tetrols for biogenic secondary organic aerosol from isoprene (Srivastava et al., 2018) (Table 1). As mineral dust tracers were considered Al, Ti, Sr, Ca, Mg (Moreno et al., 2006; Tursun et al., 2025), As, Cd, Ni indicated industrial emission (Querol et al., 2007) and EC, Sb, Sn, Fe, Pb for exhaust and non-exhaust emissions (Amato et al., 2009; Srivastava et al., 2018). Qtrue was 1574.62 and Qrobust 1574.66 which resulted in Qtrue/Qexp ratio of 1.03 which is less than 1.5 meaning the effect of outliers on the results was negligible (Belis et al., 2019a). In case of DISP test, no factor swaps were reported and the change in Q value remained minimal (-0.013). BS analysis based on 100 runs is showed in Table 2. All factors had >70 out of 100 mapped factors which is considered as appropriate solution (Belis et al., 2019a). The extra model uncertainty was 7 %. The correlation between observed and modelled PM10 concentrations was 0.92 indicating good reconstruction of PM10 mass concentration.
NaCl/COMB: long-range transport and road-salt/local combustion; PBAP: primary biological aerosol particles; BSOAIsoprene: biological secondary organic aerosol from isoprene; TRA/IND: traffic/industry factor.
Air mass transport modelling
To examine and support the physical plausibility of the resolved PMF factors that they represent realistic local and transported sources influencing the monitoring site, backward trajectories and polar plots analysis were performed. To understand airmass transport to the Ada Marina station, the standard ATMO-ACCESS Research Infrastructure (https://flexpart-request.nilu.no/data-access, last access: 8 August 2026) set-up was used for dust. Specifically, the Lagrangian particle dispersion model FLEXPART version 10.4 (Pisso et al., 2019) ran in the classic retroplume mode for 30 d backward in time for 3-hourly timesteps. The model was driven by ECMWF ERA5 reanalysis (Hersbach et al., 2020) with 0.5° spatial and hourly temporal resolution over 137 vertical levels. The resulting footprint emission sensitivity expresses the probability of potential emissions to arrive at Ada Marina, showing airmass origin, and can be converted to atmospheric concentrations at the receptor (Ada Marina) if combined with an emission dataset. For dust, emissions were first calculated using the FLEXDUST module (Groot Zwaaftink et al., 2016). FLEXDUST uses input parameters, such as soil type, vegetation/snow cover, threshold friction velocity, particle-size classes combined with surface and meteorological conditions from ECMWF ERA5 data (0.25° resolution, hourly) to calculate gridded emission fluxes for 10 size-classes of dust particles from 0.2 to 20 µm.
Beside FLEXPART, air mass transport modelling was performed also by Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) transport and dispersion model developed by the National Oceanic and Atmospheric Administration (NOAA). The 24 h back-trajectories used GFS Meteorological Data and vertical velocity model and were calculated at 100, 200 and 500 m above ground level.
The directional dependence and temporal behaviour of each factor was examined using polarPlot function from the openair (version 2.18-2), R package (version 4.4.2) (Carslaw, 2012).
Results and discussionPM<sub>10</sub> mass concentration and chemical composition
The average annual level of PM10 during WeBaSOOP campaign (23.9 µgm-3) (Table S2), was below threshold value of 40 µgm-3 (Government of the Republic of Serbia, 2013). Regulation on conditions for monitoring and requirements for air quality is based on Serbian Law on Air protection (Government of the Republic of Serbia, 2025) which is in accordance with Directive from 2008 (European Council, 2008). However, average annual level measured in Belgrade exceed the latest limit value in new EU Directive 2024/2881, that will be in force from 2030, (20 µgm-3)as well as WHOs guidelines from 2021 (15 µgm-3). PM10 concentrations showed seasonal pattern with higher levels during heating (28 µgm-3) (15 October to 15 April) than during non-heating season (19 µgm-3). Daily limit value (50 µgm-3) was exceeded four times and WHO recommended value (45 µgm-3) seven times during campaign, in heating season, with the highest daily concentration recorded on 19 December (92 µgm-3).
Carbonaceous fraction (OM + EC) was the major contributor to PM10 mass concentration with the highest contribution in winter (62 %, 20 µgm-3), then in autumn (58 %, 13.5 µgm-3), summer (49 %, 9.9 µgm-3) and spring (41 %, 8.3 µgm-3) (Fig. 2, Table S2). Organic matter was calculated as OC × 1.4 for all seasons (Turpin and Lim, 2001). The second major contributor to PM10 mass concentration was secondary inorganic aerosol (SIA) with 19 % on average. Sulphate and ammonium ions didn't show seasonal variability, while nitrate was higher in winter (2.7 µgm-3) and autumn (1.6 µgm-3) season, compared to summer (0.4 µgm-3) and spring (0.6 µgm-3). Semivolatile character of nitrate is responsible for its lower levels during warmer seasons due to increased temperatures and photochemical activity while sulphate is more stable under the same conditions (Waked et al., 2014). Mineral dust (MD) content was estimated following methodology described in Querol et al. (2001). MD contributed to PM10 mass with 13 % on annual average, without pronounced seasonality. Negligible contribution of trace elements (0.5 %, 0.1 µgm-3) and Cl− (0.6 %, 0.1 µgm-3) was observed (Fig. 2). The unaccounted PM10 mass concentration on average was 3.2 µgm-3which is around 16 %. During spring, the unaccounted PM10 mass was the highest, 4.8 µgm-3while during winter was two times lower (2.0 µgm-3). The higher fraction of unaccounted mass might be attributed to the particle-bound water content of PM10, and/or conversion factors adopted for OM and mineral dust which represent simplified calculations and may lead to uncertainties (Terzi et al., 2010).
Major composition of PM10 in Belgrade on annual level and during different seasons. Every bar represents total PM10 on annual and seasonal average concentration. Major PM10 reconstructed components were: OM: organic matter, EC: elemental carbon, SIA: secondary inorganic aerosol, MD: mineral dust, TE: trace elements.
Identification of the factors
Using the PMF model, seven factors were identified based on their chemical profiles and the tracer species characteristic of specific emission sources (Fig. 3): Biomass burning (BB); Long range transport and road-salt/local combustion (NaCl/COMB); Mineral dust (MD); Primary biological aerosol particles (PBAP); Biogenic secondary organic aerosol from isoprene (BSOAIsoprene); Ammonium sulphate; Traffic/Industry (TRA/IND). Factor fingerprints and relative contribution of the identified factors to PM10 are presented in Figs. S2 and S3. Factor chemical profiles for each factor are presented in Fig. 3, whereas time evolution of daily contributions and wind polar plots are depicted in Fig. 4.
Chemical profiles of the PMF factors in Belgrade Ada Marina. The names of the species are shown on x axis, on y axis are shown chemical concentrations of the species in ng m−3 with mean DISP values (white dots) and DISP confidence intervals (error bars). On the secondary y axis are shown relative contributions of the species apportioned by the factor (black dots). Biomass burning (BB); Long range transport and road-salt/local combustion (NaCl/COMB); Mineral dust (MD); Primary biological aerosol (PBAP); Biogenic secondary organic (BSOAIsoprene); Traffic/Industry (TRA/IND).
Daily factor contribution to PM10 mass concentration during one-year campaign (left) and wind polar plots with normalized concentration of the apportioned PMF factors (right). Biomass burning (BB); Long range transport and road-salt/local combustion (NaCl/COMB); Mineral dust (MD); Primary biological aerosol (PBAP); Biogenic secondary organic (BSOAIsoprene); Traffic/Industry (TRA/IND).
Biomass burning
Biomass burning (BB) factor contributed the most to the total PM10 concentration (21 %, Fig. S2). It was characterized by the presence of galactosan (80 % of total galactosan), 79 % mannosan, 76 % levoglucosan, 51 % of K and Cl−, 38 % of Rb and 33 % of OC (Fig. 3), typical indicators of BB. This factor also contains relatively high contribution of the As (28 %), Cd (32 %), Pb (21 %). Unlike the species that are produced during incomplete combustion (like polycyclic aromatic hydrocarbons or black carbon), heavy metals emissions from BB are related to the biomass characteristic itself (e.g., natural metal accumulation from soil) (Yao et al., 2023). Metals may also originate from surface contamination from soil and dust or activities like wood treating (e.g., with arsenic-contained preservatives) (Stojić et al., 2016).
Further examination of this factor included deriving OC / EC (9.4), OC / levoglucosan (8.3) and EC / levoglucosan ratios (0.88) which were in accordance with the values found in the literature (Table 3). Higher contribution of BB factor was observed during heating season in many European cities due to increased heating demand and stable atmospheric conditions in colder seasons (Liu et al., 2025). During cold seasons, as a result of near-surface cooling effect, temperature inversions occur, and stable boundary layer is formed which results in accumulation of the pollutants and prevention of their vertical mixing (Glojek et al., 2022). In our study, contribution of BB factor to total PM10 concentration during heating season was 34 % and temporal variability with peak during heating season (from October till April) further confirmed this factor (Fig. 4). In urban areas, BB factor is associated with domestic heating (Belis et al., 2013). In Belgrade, it is estimated that around 300 000 households are not connected to district heating, and they mostly rely on wood as heating fuel (Belgrade's air quality action plan, 2021). This indicates that residential wood burning substantially influence local air quality.
Biomass burning was a commonly identified source in Belgrade, and Balkan region as well. In the study of Stojić et al., 2016, the contribution of solid fuel burning was almost 30 % of PM10 in 2010–2011 in Belgrade. However, comparison with this study is difficult since sampling was conducted 13 years before our campaign and not all tracers needed for the complete PMF were included. The source apportionment of PM10 in Sofia, Bulgaria resolved biomass burning factor with contribution of 23 % (Hristova et al., 2020) while source apportionment study in urban area of Skopje, Macedonia, resulted in much higher contribution of BB (around 70 %) (Mirakovski et al., 2020). In case of PM2.5, contribution of BB factor was 24 % of total PM2.5 in Budapest, Hungary and 28 % in Zagreb, Croatia (Perrone et al., 2018).
OC / levoglucosan, EC / levoglucosan and OC / EC ratios from PMF BB factor, dataset and other European studies.
According to Rowell et al. (2012), wood biomass is grouped into two groups: softwood and hardwood. Levoglucosan to mannosan (L/M) ratio may be used to apportion biomass burning emissions to burning from hardwood (14–15) or softwood (<4) (Schmidl et al., 2008). The L/M ratio derived from our PMF factor was 8.9, indicating probably a mixed contribution from both soft and hardwood burning, which is further confirmed with spruce (softwood) percentage of 53 %. As shown in Fig. S4, the L/M ratios were different between summer (18.9) and winter (8.0) season suggesting possible origin of levoglucosan from other sources rather than burning of biomass for heating only. Ratios in spring and fall were consistent, 12.7 and 12.3, respectively. On 2 October, increases in levoglucosan, mannosan and galactosan concentrations were observed. The heating season in Belgrade officially begins on 15 October and if unusually low temperatures occur earlier, the season may start sooner. However, on 2 October, the ambient temperature was 15 °C which indicates another possible source of anhydrosugars beside burning of biomass for heating. Gajović and Todorović (2013) observed open fire phenomena in Serbia every year in October in the period from 2000 to 2013. The authors attributed this phenomenon to burning of agricultural waste in the autumn. During WeBaSOOP campaign, on the 2 October, satellite data showed fire occurrence in Vojvodina (Fig. S5), a common region where agricultural fires were distributed (Gajović and Todorović, 2013). The same sudden increase in biomass burning tracers' concentrations was observed in Belgrade in 2008 in study of Zangrando et al. (2016), which was also attributed to burning of agricultural waste. In their study, L/M ratio in October was in range from 2.5–16.7, which agrees with our L/M average ratio in October (14.9). Beside agricultural waste burning, the high L/M ratio in summer may also be explained by wood combustion (e.g. charcoal) in restaurants located in the Ada Ciganlija recreational area in the vicinity of Ada Marina monitoring site. The wind polar plots indicate strongest influence of air masses coming from the West in non-heating season, which coincides with direction of recreational area. Northeast, East, and Southeast wind direction may be observed during heating season, indicating influence from residential area (Fig. S6), thus supporting this interpretation.
In the previous SA studies performed for Belgrade region and Metropolitan, biomass and/or wood burning were identified as separate or mix factor. In the study of Đorđević et al. (2012) this factor was identified as separate factor based on the dominant presence of K+ (8 %) while Cvetković et al. (2015) verified it by PAHs such as Nap, Acy and Ace (10.4 %). In the studies of Todorović et al. (2020) and Almeida et al. (2020) biomass burning was identified with Cl, K, Zn, Pb and S with contribution of about 15 % in both studies. In this study biomass burning was 21 % of PM10 mass concentration, and was identified based on OC, organic tracers (levoglucosan, mannosan and galactosan) and Rb, that is, beside K, one of the alternative tracers for this factor (Massimi et al., 2020; Pereira et al., 2025).
Long-range transport and road-salt/local combustion
The next factor was a mix of NaCl from long-range transport and local emission from road salting with contribution of local combustion. It accounts for 9 % of PM10. The factor is dominated by Na (61 % of total Na), NO3- (51 %), and Cl− (30 %). As NaCl travels toward the Balkans, it is likely reacting with HNO3, forming NaNO3 and releasing HCl that is a typical aging process for sea salt aerosol (Pandolfi et al., 2020). Although Na-Cl high Pearson correlation coefficient (0.75) (Table S3) and Mg/Na (0.14) ratio (Seinfeld and Pandis, 2016) confirm marine origin, K/Na (0.22) and Ca/Na (0.18) ratios (Ooki et al., 2002) from PMF output do not comply with ratios typical for marine aerosol and thus indicate mixing with other salt sources. Extensive road salting on 20 January contributed to the pronounced peak on that day (https://nsuzivo.rs/srbija/na-auto-putevima-celu-noc-99-kamiona-prosuto-100-tona-soli-vesic-na-terenu-cak-200-radnika-veceras-cisti-saobracajnice/, last access: 8 August 2026).
A slight seasonality and peaks during the heating season (Fig. 4) point to local sources of salt from road mixed with fossil fuel and biomass combustion. This factor accounts for 13 % of EC, 11 % of levoglucosan, and 9 % of OC. Wind rose data (Fig. 4) show the dominant influence of air masses from the Northwest, aligning with the location of the “Novi Beograd” heating plant. Belgrade's district heating system covers 360 000 households and 4.4 million m2 of commercial space, with 90 % of its thermal energy derived from natural gas (https://beoelektrane.co.rs/zastita-zivotne-sredine/, last access: 8 August 2026). Although natural gas emits less particulate matter than other fuels, it remains a potential source of PM. The influence of nearby plants is supported by wind patterns corresponding to the positions of the Novi Beograd (Northwest) and, to a lesser extent, “Banovo Brdo” (Southeast) heating plants (Fig. S7). Additional fuels used by Belgrade Power Plants include CNG, light heating oil, low- and medium-sulphur heating oil, coal, and pellets which may also contribute to elevated PM10 levels during the heating season.
Flexpart footprint analysis (Fig. S8) for the two peaks in the time series of this factor shows air masses from the Atlantic and Northern Sea (Fig. S8-1) but also local influence of combustion on 19 December which is the day with the highest concentration in the whole WeBaSOOP campaign (Fig. S8-2). The NaCl/COMB factor contributed the least to PM10 mass concentration, 9 %. Similar findings were found in an urban background site in Northwest Germany, located 250 km from the coast. The aged sea salt factor in this study was explained mainly by Na, NO3-, Mg and OM and contributed to the PM10 mass same as our factor, 9 % (Beuck et al., 2011). Although located around 300 km away from closest sea, other source apportionment studies in Belgrade observed the influence of marine aerosol also. In the study of Đorđević et al. (2012), marine aerosol (14,9 %) was confirmed by the presence of Na and Cl in the coarse mode of measured particles.
Mineral dust
The mineral dust (MD) factor accounts for 17 % of PM10 and is characterized by high contributions from crustal elements: Al (78 % of total Al), Ti (69 %), Sr (62 %), Mg (60 %), and Ca (51 %), which are typical markers of soil and mineral matter. Moderate enrichments of V (30 %), Mn (29 %), and Fe (20 %) are also observed, although less pronounced than in the TRA/IND factor. Diagnostic ratio of K/Fe (0.8) complies with typical ratio in dust (Liu et al., 2022). Wind polar plots confirm the strongest influence of air masses coming from the Southwest (Fig. 4). Moreover, on 31 March 2024, high PM10 concentration (48.5 µgm-3), partially attributed to a Saharan dust outbreak, as confirmed by characteristic Na/Al (0.14) and K/Al (0.27) ratios, closely matching values reported for Saharan dust (Na/Al=0.12; K/Al=0.31) by Moreno et al. (2006). Nicolás et al. (2008) observed that Saharan outbreaks are marked by a more pronounced increase in Ti relative to Si and Ca. Therefore, Ti/Ca ratio is good marker for Saharan dust episodes. In our study, Ti/Ca (0.05) and Ti/Fe (0.14) ratios were both consistent with the values from the literature (Ti/Ca=0.02; Ti/Fe=0.10) (Nicolás et al., 2008) and aligned with ratios obtained in Debrecen, Hungary (Ti/Ca=0.07; Ti/Fe=0.09) (Borbély-Kiss et al., 2004). Although these observations were made during the March to April 2024 transition, no clear seasonal pattern was evident (Fig. 4).
Beside the long-range transported dust, the influence of local dust was also observed. The WeBaSOOP campaign coincided with preparatory work for the starting station of Belgrade metro. Sand and crushed stone were applied around the future depot site in Makiš, roughly 7 km Southwest of the Ada Marina site. This proximity is reflected in the mineral dust polar plot, which shows a joint influence of construction activity and natural mineral dust (Fig. S9).
The MD factor accounts for 17 % of PM10, but only 5 % to OC was attributed to MD, indicating no mixing of dust and organic matter by aging, and thus making it the lowest contributor to OC among all identified factors. In the study of Mijić et al. (2012), mineral matter contributed to PM10 with 19 %, with the most dominant southwest wind, which agrees with our results and indicates mineral dust constant influence in Belgrade. However, factor soil, in this study titled mineral dust, was identified in the studies of Almeida et al. (2020) and Todorović et al. (2020), contributed to PM10 mass concentration with 8 % and 5 % respectively. Our results complied with other European cities where contribution of MD factor accounted for 10 % in Milan, Italy to 25 % in Athens, Greece (Amato et al., 2016).
FLEXPART Footprint analysis for 3 h time set on 31 March confirming influence of Saharan dust on that day.
Primary biological aerosol particles
Primary biological aerosol particles (PBAP) factor contributes to 10 % of PM10 and it was traced by sugars (Glucose, Fructose, Trehalose) and sugar alcohols (Mannitol and Arabitol). It accounted for 72 % of total glucose, 68 % arabitol, 63 % trehalose, 62 % mannitol, and 57 % fructose. The remaining sugars and sugar alcohols were apportioned by PMF in the BSOAIsoprene factor. This factor shows a clear seasonal variation with higher levels during warmer seasons (Fig. 4) due to increased solar radiation and higher surface temperatures (Waked et al., 2014). PBAP factor is indicator of a highly heterogenic source of particles coming from bacterial and fungal cells or spores, their fragments (e.g. endotoxins, mycotoxins), viruses, pollen grains, and fragments of plants and insects (Barbaro et al., 2024; Samaké et al., 2019). PBAP are found mainly in the coarse fraction of PM, with a maximum in a range from 5.6 to 10 µm which can contribute to the PM mass concentration to up to 5 µgm-3 in the summer period (Marcovecchio and Perrino, 2021). As this factor gathers the particles released into the atmosphere from living organisms, they are usually referred as bioaerosols (Fröhlich-Nowoisky et al., 2016). Dispersed in the air, bioaerosols may interact with PM which then serves as a carrier for bioaerosols. As a result, allergic sensitization and respiratory infections may occur and the incidence of these diseases is higher in areas with higher air pollution (Huang et al., 2024). Mannitol and arabitol are established tracers of fungal spores while glucose is being used as tracer for plant material like pollen and fragments of plant material (Tonon et al., 2017; Samaké et al., 2019). In our study, fungal spores were estimated using mannitol to arabitol ratio (M/A) (Bauer et al., 2008). The M/A ratio (1.0) derived from the PBAP factor and dataset, aligns closely with values reported for fungal spores (Bauer et al., 2008) and for aerosol filter samples collected across Europe (Yttri et al., 2011, 2024; Samaké et al., 2019) (Table 3). Additionally, by applying specific OC to mannitol ratio of 5.2 and 10.8 to our dataset (Bauer et al., 2008), we estimated that 21 %–44 % of OCPBAP comes from fungal spores' source (Platt et al., 2026). Further resolving this heterogeneous group and estimating its contribution to the total PBAP factor remains challenging, as a broader range of primary biological organic tracers has yet to be examined (Samaké et al., 2019). The polar wind diagram shows a Northeasterly influence, likely related to the vegetation zones of the Ada Ciganlija recreational area near Ada Marina. A slight Southeasterly influence is also observed, where the Košutnjak forest is located, 3 km away and covering an area of 300 ha (Fig. S10). Borlaza et al. (2021) indicated that the influence of this source is more local than large-scale impact, complying with our conclusion that biggest influence of this factor is from vegetation at Ada Ciganlija.
The PBAP factor accounts for 10 % of PM10, 14 % to OC and 11 % to EC which was less expected result. However, in the study of Waked in Lens, France, contribution of EC to PBAP factor was 16 % on average annual level. Such results were obtained due to possible atmospheric mixing process, limitations in PMF modelling or OC / EC analysis artifacts. Possible explanation for this may be that some types of PBAP, char and evolve as a modern carbon EC during OC / EC analysis (Yttri et al., 2021). Similarly to our study, PBAP contributions at two urban sites in France were 9 % Lens (Waked et al., 2014) and 11 % in Grenoble, (Srivastava et al., 2018). Slightly higher contributions (16 %) were measured in a peri-urban area 25 km away from Rome, Italy (Marcovecchio and Perrino, 2021).
Mannitol to arabitol ratios for PBAP factor from both PMF and dataset compared with other studies in Europe.
a Vienna, Austria; b Oslo, Norway, summer and winter season, respectively; c Zeppelin Observatory, Birkenes Observatory, respectively; d France.
Biogenic secondary organic aerosol from isoprene
This factor represents biogenic secondary organic aerosol formed from isoprene oxidation (BSOAIsoprene), a major volatile organic compound emitted from land vegetation (van Drooge and Grimalt, 2015). It accounts for 10 % of PM10 (Fig. 3) and is characterized by the presence of 2-methylerythritol (84 % of total 2-methylerythritol) and 2-methylthreitol (76 %), both known isoprene oxidation products (Claeys et al., 2004). A modest contribution from PBAP tracers (17 %–28 % of their total content) likely reflects co-varying biological activity, as these two factors exhibit similar temporal patterns (Fig. 4), although these species do not define this factor. Liu et al. (2025) emphasised that including secondary biogenic organic aerosol tracers, such as 2-methyltetrols or pinic acid and 3-MBTCA (markers for α-pinene oxidation), in PMF analysis led to the separation of a substantial OC fraction in the form of biogenic secondary organic aerosol from a sulphate-rich factor. A similar split was observed in our study. This is supported by the co-occurrence of SO42− (24 %) and NH4+ (22 %), present as NH4HSO4 and/or (NH4)2SO4, suggesting that 2-methyltetrols are formed via acid-catalysed aqueous-phase reactions, confirming the secondary origin of this factor. Wind polar plot shows influence from East, and both North and Southeast, indicating influence from the nearby vegetational sources as well as influence from Košutnjak forest, even higher than for PBAP factor (Fig. 4). The BSOAIsoprene factor appears exclusively in the growing season, peaking in summer. It serves as a clear proxy for isoprene-derived SOA formed under conditions of high photochemical activity and enhanced biogenic emissions. It contributes 10 % to the PM10 mass concentration and 11 % to OC. Similarly to our result, average annual contribution to PM10of biogenic secondary organic aerosol was in range from 8 %–11 % in the study of Borlaza et al. (2021) in Grenoble, France. The similar contribution to OC of 10 % was observed in Norway (Yttri et al., 2021). To date, to our knowledge, no source apportionment study including tracers for secondary biogenic aerosol particles has been conducted in Belgrade, except to other WeBaSOOP campaign papers.
Ammonium sulphate
This ammonium sulphate is a secondary factor accounting for 18 % of PM10 and represents the long-range transport of aged air masses and secondary aerosol formation. It is dominated by NH4+ (55 % of total NH4+) and SO42- (54 %), with the coexistence of NH4HSO4 and (NH4)2SO4 indicating only partial neutralization of H2SO4. An elevated NO3- fraction (17 %) further supports its secondary origin. Trace contributions from As (18 %), Cd (14 %), and levoglucosan (5 %) point to minor influences from industrial and combustion sources, likely associated with long-range transport. A high OC / EC ratio (9.1) corroborates the secondary and aged nature of this factor.
The occurrence of NH4HSO4 and (NH4)2SO4 in both, this and in the BSOAIsoprene factor likely reflects a common regional background that persists year-round but co-varies with distinct seasonal processes. In summer, partially neutralized sulfuric acid acts as an acid seed facilitating IEPOX (Isoprene Epoxydiol) uptake and subsequent 2-methyltetrol formation. During other seasons, it primarily represents regional SIA. The (NH4)2SO4 factor accounts for 18 % of PM10 and 12 % of OC, making it the second largest contributor to PM10 after the BB factor. The contribution of this factor is similar to the ones previously reported for various urban sites in Europe (Beuck et al., 2011; Borlaza et al., 2021) as well as in Belgrade. Đorđević et al. (2012) found that the formation of (NH4)2SO4 has the dominant influence on the contents of aerosol ionic species in Belgrade in 2008. Belis et al. (2019b) indicated that secondary sulphate in the Danube and Western Balkan region is mostly associated with coal combustion since its precursor is gas SO2. In the study of Todorović et al. (2020), factors secondary sulphate from regional combustion sources (S, NH4+) (28.8 %) and local combustion sources and ammonium nitrate (V, Ni, NO3-, NH4+) (29.7 %) accounted for around 2/3 of PM10 total mass concentration. For the same sampling site, Almeida et al. (2020) identified sulphate factor with contribution of 3 %, which was separated from oil fuel combustion (38 %). Comparing their results to ammonium sulphate contribution in our study (18 %), differences may be explained by nonidentical PM fraction, SA with PM10 identified factors that are not only from combustion sources, the chemicals analysed and used as input data in PMF, time period when studies were conducted as well as long-term trends that indicate airborne PM decrease (De Hoogh et al., 2025). Almeida et al. (2020) analysed 19 elements (Table S1) while Todorović et al. (2020) analysed ions beside the 19 elements (Table S1). Furthermore, Todorović et al. (2020) as well as Almeida et al. (2020) had sampling campaign in the period from 2014–2015. Their studies represent the last SA studies in Belgrade until WeBaSOOP sampling campaign. In the time span of eight years, the sources in Belgrade changed due to: substantial changes in traffic routes in Belgrade central zone, industrial activities, observed reductions in PM10 concentrations during the 2010s. The PM10 decrease likely reflect the combined effects of the long-term replacement of coal- and oil-fired boiler houses with district heating supplied predominantly by natural gas, together with the gradual phase-out of small local boiler plants and other emission control measures (https://beoelektrane.co.rs/saopstenja/veliki-doprinos-beogradskih-elektrana-u-borbi-za-cistiji-vazduh-i-smanjenje-zagadenja-u-glavnom-gradu/, last access: 8 August 2026). Finally, Almeida et al. (2020) emphasised that in their study contributions of some tracers were not well defined which led to overestimation of oil combustion and underestimation of other factors.
Wind polar plots showed highest influence from the Southeast and South (Fig. 4). The “Kolubara” thermal power plant is located around 50 km South of Belgrade, and “Kostolac” is approximately 60 km East. Peak on 14 October showed HYSPLIT backward trajectories from the Southwest and influence on that day may be attributed to operation of “Nikola Tesla A” thermal power plant, located around 25 km way from Belgrade (Fig. S11). Our results complied with results obtained previously in Belgrade. In the study of Stojić et al. (2016), secondary aerosol contributed to PM10 19.5 %, while contribution of 16 % was obtained in Sofia, Bulgaria (Hristova et al., 2020). Our results are withing the European range too, 15 % to 26 % (Liu et al., 2025).
Traffic/industry
This Traffic/Industry factor (TRA/IND) is likely a mixture of tail pipe and non-tail pipe emissions with influence of industrial activity based on presence of As, Cd, Pb and V. It accounts for 15 % of PM10 and it is characterized by EC (41 % of total EC) and a broad suite range of trace metals, notably Fe (64 %), Sb (60 %), Cu (57 %), Sn (57 %), Mn (51 %), Pb (37 %), V (35 %), Cd (32 %), and As (25 %). Fe, Sb, Sn and Cu are typical non-exhaust tracers (e.g., from brake wear), V relates to heavy oil combustion and As, Cd and Pb are associated with coal combustion and metallurgic activities (Yttri et al., 2021; Pandolfi et al., 2016). The enrichment of Ca (34 %) and Mg (22 %) suggests road dust resuspension and influence of construction activity. Therefore, this factor is considered as a mix of traffic exhaust and non-exhaust emissions and industrial sources. Low contributions from levoglucosan (6 %), 2-methyltetrols (<0.5 %), and sugars and sugar alcohols (<5 %) confirm a minor to negligible influence by biomass burning, biogenic SOA, and PBAP. SIA species are moderate for NO3- (12 %), minor for SO42- (2 %) and absent for NH4+, consistent with limited regional influence. This factor does not change throughout a year. The TRA/IND factor contributed 15 % of PM10 and 17 % of OC, making it the second largest contributor to OC. The OC / EC ratio of 2.5 is higher than the typical OC / EC traffic ratio (1.4) found in literature (Pio et al., 2011), which indicates the contribution of other combustion sources to this factor (probably SOA mixed with primary emissions). Since the PMF model results are based on internal correlations among species that have similar time series, it is difficult to differentiate sources that do not vary independently. The limitation of 24 h sampling methodology is that it couldn't capture the diurnal variations of traffic source, therefore, some primary and secondary sources are not separated (Waked et al., 2014).
Traffic with metal industry influence was also observed in Rijeka, Croatia and accounted for 11 % of total PM10 mass concentration (Mifka et al., 2021). In other countries of the Balkan region, traffic contributed 9 % in Sofia, Bulgaria, and 18.9 % in Istanbul, Turkey, with slight seasonality observed in winter (Hristova et al., 2020; Koçak et al., 2011). Our results also complied with range observed in other European urban background sites (5 %–25 %) (Srivastava et al., 2018).
Conclusion
This study presents the chemical composition and source apportionment of PM10 mass concentrations during the WeBaSOOP campaign (June 2023 to May 2024). Although average PM10 concentration complied with current EU and national legislation, 23.9 µgm-3, it exceeded the limit value proposed by the new EU 2024/2881 AQ Directive on ambient air quality and cleaner air for Europe, to be in force from 2030, and the WHO AQ Guidelines updated in 2021. By the application of mass reconstruction model, it was estimated that carbonaceous fraction (OM + EC) constituted the dominant PM component (54 % on annual average), followed by secondary inorganic aerosol (19 %) and mineral dust (13 %). Source apportionment using USEPA PMF 5.0 resolved seven factors contributing to PM10 as follows: biomass burning (21 %), ammonium sulphate (18 %), mineral dust (17 %), traffic/industry (15 %), primary biological aerosol particles (10 %), biogenic secondary organic aerosol from isoprene (10 %), and long-range transport and road salt/local combustion (9 %).
Eight SA studies were conducted in the past in Belgrade urban and Metropolitan area upon data sets collected in the period from 2003 and 2015 and WeBaSOOP year-long campaign in 2023–2024. Although WeBaSOOP study was conducted upon data set collected eight years after the last study and having different, broader range of, species used for SA, factors traffic, industry, different combustion sources, mineral/soil dust and secondary inorganic aerosol were identified in all studies in the past and ours. This indicates a continuity of these factors throughout the years on air quality in Belgrade. One study performed on data sets collected 15 years ago identified marine aerosol also but with higher contribution than in our, WeBaSOOP, study. However, in contrast to other studies, this was the first one that identified factors from biological origin (PBAP and BSOAIsoprene).
Evaluation of ratios of PM elements and species, back trajectory analysis and wind polar plots verified the local and regional sources contributing to PM10 in Belgrade. Compatibility of the two approaches, mass reconstruction and SA, indicated that more than 40 % of PM10 mass was related to different combustion sources, between 30 %–35 % may be devoted to mineral dusts and inorganic aerosols of anthropogenic and natural sources, while specific organic tracers proved that about 20 % of PM10 mass originated from biological sources. Seasonal patterns varied among factors. Biomass burning and the long-range transport and road salt/local combustion factors exhibited elevated levels during heating season. The biomass burning driven primarily by residential wood burning and the long-range transport and road salt/local combustion reflecting contributions from local sources during winter period of road salt events and local combustion. In contrast, primary biological aerosol particles and biogenic secondary organic aerosol from isoprene peaked during the non-heating season, coinciding with increased biological activity. Mineral dust, ammonium sulphate, and traffic/industry showed no clear seasonal variability. Ammonium sulphate was resolved as a distinct factor, but its influence was also evident in the BSOAIsoprene factor, supporting the secondary origin of this aerosol and its likely role in facilitating isoprene oxidation. Traffic factor has more enhanced diurnal variation than seasonal which cannot be observed with resolution of 24 h sampling. Therefore, time series of traffic and industry tracers were not different enough to separate them which is a limitation of this 24 h sampling resolution.
Incorporation of specific organic tracers in PMF enabled the identification of biogenic activity-related factors and provided deeper insight into the composition and sources of PM10 in the Belgrade Metropolitan area. Our results showed influence of both, local (biomass burning, traffic, biogenic aerosols) and regional sources (mineral dust, salt from long-range transport) on PM10 concentration in this area. Continuous measurements and source apportionment are needed in data-scarce areas of Europe to better understand source impacts on air quality. Therefore, such findings may offer a scientific basis for air quality management and support policy makers in identifying priority sectors and targeted measures to mitigate PM10 pollution. However, further work for better understanding of PM10 regional and local sources in Belgrade is needed by performing SA study at least for two fractions of airborne PM represent by fine and coarse fraction as well as combination of offline and online PMF analysis and long-term measurements.
Mineral dust events, marine aerosol and biogenic aerosols contribute to the background concentration of PM10 and PM2.5 and are largely beyond the scope of emission control measures. Although sea salt was not extracted as a separate factor in WeBaSOOP study, its contribution to PM10 mass may be accounted as 1–2 µgm-3, but in case of mineral dust, contribution is estimated to be around 4 µgm-3. In WeBaSOOP study, mineral dust represents mix of both natural and anthropogenic sources from local and long-range transport, while 1/5 of PM10 mass (around 5 µgm-3) was attributed to primary and secondary airborne particulate matter of biogenic sources. As emissions from these sources are beyond controllable measures compliance with ambient air quality limit values consequently requires sufficient reductions in controllable anthropogenic emissions to offset the irreducible natural background. Where legislation permits, documented contributions from specific natural sources, such Sahara dust events and sea salt, may be excluded from compliance assessment. However, routine exclusion of biogenic aerosol contributions is generally not possible to be applied without measurements such WeBaSOOP study enabled.
Data availability
All data on OC, EC, ions, elements, and organic tracers are available at EBAS (https://ebas.nilu.no/, last access: 8 August 2026): 10.48597/J32C-ZYJW (Jovasevic-Stojanovic et al., 2024c), 10.48597/5JNE-638V (Jovasevic-Stojanovic et al., 2024a), 10.48597/8TQV-HXDA (Jovasevic-Stojanovic et al., 2024b), 10.48597/AJ4A-EZ64 (Yttri, 2024).
The supplement related to this article is available online at https://doi.org/10.5194/acp-26-14165-2026-supplement.
The contact author has declared that none of the authors has any competing interests.
Disclaimer
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. The authors bear the ultimate responsibility for providing appropriate place names. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
Acknowledgements
For participation in data collection, the authors acknowledge the team from Vinča Institute and Mining and Metallurgy Institute Bor, which participated in the WeBaSOOP campaign, and the WeBaSOOP team from the Public Health Institute of Belgrade for their logistical support and valuable contributions before and during the experimental campaign. FLEXPART results used a virtual access service that is supported by the European Commission under the Horizon 2020 – Research and Innovation Framework Programme, H2020-INFRAIA-2020-1, ATMO-ACCESS. GA No 101008004. The computations/simulations were performed using resources provided by Sigma2 – the National Infrastructure for High-Performance Computing and Data Storage in Norway. The authors acknowledge Rosa Lara Bueno and Joaquim Cortes Hurtado from IDAEA CSIC for their contribution in visualisation of chemical profiles including uncertainties and Nikolaos Evangeliou form NILU for contribution in FLEXPART air transport modelling analysis. AI tools were used for language editing and clarity improvement.
Financial support
This research has been supported by the European Commission, EU HORIZON EUROPE Framework Programme (grant no. 101060170) and the Ministry of science, technological development and innovation (grant no. 451-03-33/2026-03/ 200017).
Review statement
This paper was edited by Benjamin A Nault and reviewed by two anonymous referees.
ReferencesAas, W., Fagerli, H., Alastuey, A., Cavalli, F., Degórska, A., Feigenspan, S., Brenna, H., Gliß, J., Heinesen, D., Hueglin, C., Holubová, A., Jaffrezo, J.-L., Mortier, A., Murovec, M., Putaud, J.-P., Rüdiger, J., Simpson, D., Solberg, S., Tsyro, S., Tørseth, K., and Yttri, K. E.: Trends in air pollution in Europe, 2000–2019, Aerosol Air Qual. Res., 24, 230237, 10.4209/aaqr.230237, 2024.Almeida, S., M., Manousakas, M., Diapouli, E., Kertesz, Z., Samek, L., Hristova, E., Šega, K., Padilla Alvarez, R., Belis, C. A., Eleftheriadis, K., and The IAEA European Region Study Group: Ambient particulate matter source apportionment using receptor modelling in European and Central Asia urban areas, Environ. Pollut., 266, 115199, 10.1016/j.envpol.2020.115199, 2020.Amato, F., Pandolfi, M., Viana, M., Querol, X., Alastuey, A., and Moreno, T.: Spatial and chemical patterns of PM10 in road dust deposited in urban environment, Atmos. Environ., 43, 1650–1659, 10.1016/j.atmosenv.2008.12.009, 2009.Amato, F., Alastuey, A., Karanasiou, A., Lucarelli, F., Nava, S., Calzolai, G., Severi, M., Becagli, S., Gianelle, V. L., Colombi, C., Alves, C., Custódio, D., Nunes, T., Cerqueira, M., Pio, C., Eleftheriadis, K., Diapouli, E., Reche, C., Minguillón, M. C., Manousakas, M.-I., Maggos, T., Vratolis, S., Harrison, R. M., and Querol, X.: AIRUSE-LIFE+: a harmonized PM speciation and source apportionment in five southern European cities, Atmos. Chem. Phys., 16, 3289–3309, 10.5194/acp-16-3289-2016, 2016.Barbaro, E., Feltracco, M., De Blasi, F., Turetta, C., Radaelli, M., Cairns, W., Cozzi, G., Mazzi, G., Casula, M., Gabrieli, J., Barbante, C., and Gambaro, A.: Chemical characterization of atmospheric aerosols at a high-altitude mountain site: a study of source apportionment, Atmos. Chem. Phys., 24, 2821–2835, 10.5194/acp-24-2821-2024, 2024.Bauer, H., Claeys, M., Vermeylen, R., Schueller, E., Weinke, G., Berger, A., and Puxbaum, H.: Arabitol and mannitol as tracers for the quantification of airborne fungal spores, Atmos. Environ., 42, 588–593, 10.1016/j.atmosenv.2007.10.013, 2008.Belgrade's air quality action plan: Belgrade's air quality action plan (2021–2031), The Official Gazette of Belgrade, 46/2021 http://demo.paragraf.rs/demo/combined/Old/t/t2021_07/BG_046_2021_001.htm (last access: 8 August 2026), 2021 (in Serbian).Belis, C. A., Karagulian, F., Larsen, B. R., and Hopke, P. K.: Critical review and meta-analysis of ambient particulate matter source apportionment using receptor models in Europe, Atmos. Environ., 69, 94–108, 10.1016/j.atmosenv.2012.11.009, 2013.Belis, C. A., Favez, O., Mircea, M., Diapouli, E., Manousakas M.-I., Vratolis, S., Gilardoni, S., Paglione, M., Decesari, S., Mocnik, G., Mooibroek, D., Salvador, P., Takahama, S., Vecchi, R., and Paatero P.: European guide on air pollution source apportionment with receptor models – Revised version 2019, EUR 29816 EN, Publications Office of the European Union, Luxembourg, https://data.europa.eu/doi/10.2760/439106 (last access: 8 August 2026), 2019a.Belis, C. A., Pisoni, E., Degraeuwe, B., Peduzzi, E., Thunis, P., Monforti-Ferrario, F., and Guizzardi, D.: Urban pollution in the Danube and Western Balkans regions: The impact of major PM2.5 sources, Environ. Int., 133, 105158, 10.1016/j.envint.2019.105158, 2019b.Beuck, H., Quass, U., Klemm, O., and Kuhlbusch, T. A. J.: Assessment of sea salt and mineral dust contributions to PM10 in NW Germany using tracer models and positive matrix factorization, Atmos. Environ., 45, 5813–5821, 10.1016/j.atmosenv.2011.07.010, 2011.Borbély-Kiss, I., Kiss, A. Z., Koltay, E., Szabó, G., and Bozó, L.: Saharan dust episodes in Hungarian aerosol: elemental signatures and transport trajectories, J. Aerosol Sci., 35, 1205–1224, 10.1016/j.jaerosci.2004.05.001, 2004.Borlaza, L. J. S., Weber, S., Uzu, G., Jacob, V., Cañete, T., Micallef, S., Trébuchon, C., Slama, R., Favez, O., and Jaffrezo, J.-L.: Disparities in particulate matter (PM10) origins and oxidative potential at a city scale (Grenoble, France) – Part 1: Source apportionment at three neighbouring sites, Atmos. Chem. Phys., 21, 5415–5437, 10.5194/acp-21-5415-2021, 2021.Bozzetti, C., Daellenbach, K, R., Hueglin, C., Fermo, P., Sciare, J., Kasper-Giebl, A., Mazar, Y., Abbaszade, G., El Kazzi, M., Gonzalez, R., Shuster-Mieseles, T., Flasch, M., Wolf, R., Krepelova, A., Canonaco, F., Schnelle-Kreis, J., Slowik, J. G., Zimmermann, R., Rudich, Y., Baltensperger, U., El Haddad, I., and Prevot, A. S. H.: Size-resolved identification, characterization, and quantification of primary biological organic aerosol at a European rural site, Environ. Sci. Technol., 50, 3425–3434, 10.1021/acs.est.5b05960, 2016.Burnett, R., Chen, H., Szyszkowicz, M., Fann, N., Hubbell, B., Pope III, C. A., Apte, J. S., Brauer, M., Cohen, A., Weichenthal, S., Coggins, J., Di, Q., Brunekreef, B., Frostad, J., Lim, S. S., Kan, H., Walker, K. D., Thurston, G. D., Hayes, R. B., Lim, C. C., Turner, M. C., Jerrett, M., Krewski, D., Gapstur, S. M., Diver, W. R., Ostro, B., Goldberg, D., Crouse, D. L., Martin, R. V., Peters, P., Pinault, L., Tjepkema, M., van Donkelaar, A., Villeneuve, P. J., Miller, A. B., Yin, P., Zhou, M., Wang, L., Janssen, N. A. H., Marra, M., Atkinson, R. W., Tsang, H., Thach, T. Q., Kan, H., Cannon, J. B., Allen, R. T., Hart, J. E., Laden, F., Cesaroni, G., Forastiere, F., Weinmayr, G., Jaensch, A., Nagel, G., Concin, H., and Spadaro, J. V.: Global estimates of mortality associated with long-term exposure to outdoor fine particulate matter, P. Natl. Acad. Sci. USA, 115, 9592–9597, 10.1073/pnas.1803222115, 2018.Cavalli, F., Viana, M., Yttri, K. E., Genberg, J., and Putaud, J.-P.: Toward a standardised thermal-optical protocol for measuring atmospheric organic and elemental carbon: the EUSAAR protocol, Atmos. Meas. Tech., 3, 79–89, 10.5194/amt-3-79-2010, 2010.Carslaw, D. C.: The Openair Manual – Open-Source Tools for Analysing Air Pollution Data, Manual for version 0.6-0, University of York, https://cran.r-project.org/web/packages/openair/openair.pdf (last access: 8 August 2026), 2012.Choi, Y., Byun, G., Kim, H., Stewart, R., Song, Y., Heo, S., Lee, J.-T., Tong, S., Lavigne, E., Valdés Ortega, N., Matus Correa, P., Osorio, S., Achilleos, S., Kyselý, J., Urban, A., Roye, D., Orru, H., Maasikmets, M., Jaakkola, J. J. K., Ryti, N., Pascal, M., Schneider, A., Breitner, S., Katsouyanni, K., Samoli, E., Krage Carlsen, H., Entezari, A., Mayvaneh, F., Raz, R., Stafoggia, M., de'Donato, F., Hashizume, M., Ng, C. F. S., Madaniyazi, L., Hurtado Diaz, M., Félix Arellano, E. E., Klompmaker, J., Rao, S., Madureira, J., Gaio, V., Guo, Y., Scovronick, N., Garland, R. M., Kim, H., Lee, W., Forsberg, B., Vicedo-Cabrera, A. M., Ragettli, M. S., Guo, Y. L., Pan, S.-C., Armstrong, B., Sera, F., Gasparrini, A., Masselot, P., Mistry, M., Zanobetti, A., Schwartz, J., and Bell, M. L.: Temporal changes in mortality risk associated with PM10 across 143 cities in 26 countries: a multicountry, multicity time-series study, Lancet Planet. Health, 10, 101465, 10.1016/j.lanplh.2026.101465, 2026.Ćirović, Ž., Stojanović, D. B., Davidović, M., Onjia, A., Garcia-Marlès, M., Lozano, N. P., Alastuey, A., and Jovašević-Stojanović, M.: Concentrations and Estimation of Sources of Ultrafine Particles in the City of Belgrade at Ada Marina Urban Background Site, Environments, 13, 47, 10.3390/environments13010047, 2026.Claeys, M., Graham, B., Vas, G., Wang, W., Vermeylen, R., Pashynska, V., Cafmeyer, J., Guyon, P., Andreae, M. O., Artaxo, P., and Maenhaut, W.: Formation of secondary organic aerosols through photooxidation of isoprene, Science, 303, 1173–1176, 10.1126/science.1092805, 2004.Cohen, A. J., Brauer, M., Burnett, R., Anderson, H. R., Frostad, J., Estep, K., Balakrishnan, K., Brunekreef, B., Dandona, L., Dandona, R., Feigin, V., Freedman, G., Hubbell, B., Jobling, A., Kan, H., Knibbs, L., Liu, Y., Martin, R., Morawska, L., Pope III, C. A., Shin, H., Straif, K., Shaddick, G., Thomas, M., van Dingenen, R., van Donkelaar, A., Vos, T., Murray, C. J. L., and Forouzanfar, M. H.: Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: an analysis of data from the Global Burden of Diseases Study 2015, The Lancet, 389, 1907–1918, 10.1016/S0140-6736(17)30505-6, 2017.Cvetković, A., Jovašević-Stojanović, M., Marković, D., and Ristovski, Z.: Concentration and source identification of polycyclic aromatic hydrocarbons in the metropolitan area of Belgrade, Serbia, Atmos. Environ., 112, 335–343, 10.1016/j.atmosenv.2015.04.034, 2015.De Hoogh, K., Hoek, G., Flückiger, B., Bussalleu, A., Vienneau, D., Jeong, A., Probst-Hensch, N., de Pinho, M. G. M., Mackenbach, J. D., Lakerveld, J., Beulens, J. W. J., Castagné, R., Delpierre, C., Kelly-Irving, M., Shen, Y., Huss, A., Dadvand, P., Cirach Pradas, M., Nieuwenhuijsen, M., Vlaanderen, J., and Vermeulen, R.: A Europe-wide characterization of the external exposome: A spatio temporal analysis, Environ. Int., 200, 109542, 10.1016/j.envint.2025.109542, 2025.Đorđević, D., Mihajlidi-Zelić, A., Relić, D., Ignjatović, L., Huremović, J., Stortini, A. M., and Gambaro, A.: Size-segregated mass concentration and water soluble inorganic ions in an urban aerosol of the Central Balkans (Belgrade), Atmos. Environ., 46, 309–317, 10.1016/j.atmosenv.2011.09.057, 2012.European Council: On Ambient Air Quality and Cleaner Air for Europe 2008/50/EC, Off. J. Eur. Union, 1, 1–44, https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32008L0050&from=en (last access: 8 August 2026), 2008.European Council: On Ambient Air Quality and Cleaner Air for Europe 2024/2881, Off. J. Eur. Union, 1, 1–70, https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=OJ:L_202402881 (last access: 8 August 2026), 2024.European Environment Agency: Harm to human health from air pollution in Europe: burden of disease status, https://www.eea.europa.eu/en/analysis/publications/harm-to-human-health-from-air-pollution-burden-of-disease-status-2025 (last access: 4 December 2025), 2025a.European Environment Agency: Air quality status report 2025: Particulate matter – PM2.5/ PM10, Publications Office of the European Union, https://www.eea.europa.eu/en/analysis/publications/air-quality-status-report-2025/particulate-matter (last access: 4 December 2025), 2025b.Fröhlich-Nowoisky, J., Kampf, C. J., Weber, B., Huffman, J. A., Pöhlker, C., Andreae M. O., Lang-Yona, N., Burrows, S. M, Gunthe, S. S., Elbert, W., Su, H., Hoor, P., Thines, E., Hoffmann, T., and Després, V. R.: Bioaerosols in the Earth system: Climate, health, and ecosystem interactions, Atmos. Res., 182, 346–376, 10.1016/j.atmosres.2016.07.018, 2016.Gajović, V. and Todorović, B.: Spatial and temporal analysis of fires in Serbia for period 2000–2013, J. Geogr. Inst. Jovan Cvijić SASA, 63, 297–312, 10.2298/IJGI1303297G, 2013.Gilardoni, S., Vignati, E., Cavalli, F., Putaud, J. P., Larsen, B. R., Karl, M., Stenström, K., Genberg, J., Henne, S., and Dentener, F.: Better constraints on sources of carbonaceous aerosols using a combined 14C – macro tracer analysis in a European rural background site, Atmos. Chem. Phys., 11, 5685–5700, 10.5194/acp-11-5685-2011, 2011.Glojek, K., Močnik, G., Alas, H. D. C., Cuesta-Mosquera, A., Drinovec, L., Gregorič, A., Ogrin, M., Weinhold, K., Ježek, I., Müller, T., Rigler, M., Remškar, M., van Pinxteren, D., Herrmann, H., Ristorini, M., Merkel, M., Markelj, M., and Wiedensohler, A.: The impact of temperature inversions on black carbon and particle mass concentrations in a mountainous area, Atmos. Chem. Phys., 22, 5577–5601, 10.5194/acp-22-5577-2022, 2022.Government of the Republic of Serbia: Regulation on conditions for monitoring and air quality requirements, Official Gazette of Republic of Serbia, 63, https://www.paragraf.rs/propisi/uredba-uslovima-monitoring-zahtevima-kvaliteta-vazduha.html (last access: 8 August 2026), 2013.Government of the Republic of Serbia: Law on air protection, Official Gazette of Republic of Serbia, 51, https://www.paragraf.rs/propisi/zakon_o_zastiti_vazduha.html (last access: 8 August 2026), 2025.Groot Zwaaftink, C. D., Grythe, H., Skov, H., and Stohl, A.: Substantial contribution of northern high-latitude sources to mineral dust in the Arctic. J. Geophys. Res.-Atmos., 121, 678–697, 10.1002/2016JD025482, 2016.Health Effects Institute: Trends in Air Quality and Health in the Republic of Serbia: A State of Global Air Special Report. Boston, MA, Health Effects Institute, https://www.stateofglobalair.org/sites/default/files/documents/2022-09/soga-southeast-europe-serbia-report-english.pdf (last access: 8 August 2026), 2022.Health Effects Institute: State of Global Air 2024, Special Report, Boston, MA:Health Effects Institute, https://www.stateofglobalair.org/resources/archived/state-global-air-report-2024 (last access: 8 August 2026), 2024.Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolás, J., Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., De Chiara, G., Dahlgren, P., Dee, D., Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer, A., Haimberger, L., Healy, S., Hogan, R. J., Hólm, E., Janisková, M., Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., de Rosnay, P., Rozum, I., Vamborg, F., Villaume, S., and Thépaut, J.-N.: The ERA5 global reanalysis, Q. J. Roy. Meteor. Soc., 146, 1999–2049, 10.1002/qj.3803, 2020.Hristova, E., Veleva, B., Georgieva, E., and Branzov, H.: Application of positive matrix factorization receptor model for source identification of PM10 in the city of Sofia, Bulgaria, Atmos., 11, 890, 10.3390/atmos11090890, 2020.Huang, Z., Yu, X., Liu, Q., Maki, T., Alam, K., Wang, Y., Xue, F., Tang, S., Du, P., Dong, Q., Wang, D., and Huang, J.: Bioaerosols in the atmosphere: A comprehensive review on detection methods, concentration and influencing factors, Sci. Total Environ., 912, 168818, 10.1016/j.scitotenv.2023.168818, 2024.IARC: IARC monographs on the evaluation of carcinogenic risks to humans Volume 109 Outdoor air pollution, International Agency for Research on Cancer, Lyon, France, https://publications.iarc.who.int/Book-And-Report-Series/Iarc-Monographs-On-The-Identification-Of-Carcinogenic-Hazards-To-Humans/Outdoor-Air-Pollution-2015 (last access: 8 August 2026), 2016.Ivezić, D., Živković, M., Danilović, D., Madžarević, A., and Tanasijević, M.: The state and perspective of the natural gas sector in Serbia, Energ. Source Part B, 11, 1061–1067, 10.1080/15567249.2013.858796, 2016.Jovasevic-Stojanovic, M., Kovacevic, R., and Radovic, B.: Inorganics in air and particle phase at Beograd Ada Marina, EBAS [data set], 10.48597/5JNE-638V (last access: 8 August 2026), 2024a.Jovasevic-Stojanovic, M., Kovacevic, R., and Radovic, B.: Heavy metals and inorganics in air and particle phase at Beograd Ada Marina, EBAS [data set], 10.48597/8TQV-HXDA (last access: 8 August 2026), 2024b.Jovasevic-Stojanovic, M., Petrovic, B., and Davidovic, M.: OC/EC at Beograd Ada Marina, EBAS [data set], 10.48597/J32C-ZYJW (last access: 8 August 2026), 2024c.Khan, S., Gurjar, B. R., and Sahu, V.: Deposition modeling of ambient particulate matter in the human respiratory tract, Atmos. Pollut. Res., 13, 101565, 10.1016/j.apr.2022.101565, 2022.Koçak, M., Theodosi, C., Zarmpas, P., Im, U., Bougiatioti, A., Yenigun, O., and Mihalopoulos, N.: Particulate matter (PM10) in Istanbul: Origin, source areas and potential impact on surrounding regions, Atmos. Environ., 45, 6891–6900, 10.1016/j.atmosenv.2010.10.007, 2011.Liu, X., Turner, J. R., Hand, J. L., Schichtel, B. A., and Martin, R. V.: A global‐scale mineral dust equation, J. Geophys. Res.-Atmos., 127, e2022JD036937, 10.1029/2022JD036937, 2022.Liu, X., Zhang, X., Jin, B., Wang, T., Qian, S., Zou, J., Dinh, V. N. T., Jaffrezo, J. L., Uzu, G., Dominutti, P., Darfeuil, S., Favez, O., Conil, S., Marchand, N., Castillo, S., de la Rosa, J. D., Grange, S., Hueglin, C., Eleftheriadis, K., Diapouli, E., Manousakas, M. I., Gini, M., Nava, S., Calzolai, G., Alves, C., Monge, M., Reche, C., Harrison, R. M., Hopke, P. K., and Querol, X.: Source apportionment of PM10 based on offline chemical speciation data at 24 European sites, npj Clim. Atmos. Sci., 8, 255, 10.1038/s41612-025-01097-7, 2025.Marcovecchio, F. and Perrino, C.: Contribution of primary biological aerosol particles to airborne particulate matter in indoor and outdoor environments, Chemosphere, 264, 128510, 10.1016/j.chemosphere.2020.128510, 2021.Massimi, L., Simonetti, G., Buiarelli, F., Di Filippo, P., Pomata, D., Riccardi, C., Ristorini, M., Astolfi, M. L., and Canepari, S.: Spatial distribution of levoglucosan and alternative biomass burning tracers in atmospheric aerosols in an urban and industrial hot-spot of Central Italy, Atmos. Environ., 239, 117740, 10.1016/j.atmosres.2020.104904, 2020.McDuffie, E. E., Martin, R. V., Spadaro, J. V., Burnett, R., Smith, S. J., O'Rourke, P., Hammer, M. S., van Donkelaar, A., Bindle, L., Shah, V., Jaeglé, L., Luo, G., Yu, F., Adeniran, J. A., Lin, J., and Brauer, M.: Source sector and fuel contributions to ambient PM2.5 and attributable mortality across multiple spatial scales, Nat. Commun., 12, 3594, 10.1038/s41467-021-23853-y, 2021.Mifka, B., Žurga, P., Kontošić, D., Odorčić, D., Mezlar, M., Merico, E., Grasso, F. M., Conte, M., Contini, D., and Alebić-Juretić, A.: Characterization of airborne particulate fractions from the port city of Rijeka, Croatia, Mar. Pollut. Bull., 166, 112236, 10.1016/j.marpolbul.2021.112236, 2021.Mijailović, R., Marković, N., Pešić, D., and Vlajić, J. V.: Evaluation of scenarios for improving energy efficiency and reducing exhaust emissions of a passenger car fleet: A methodology, Transp. Res. D, 73, 352–366, 10.1016/j.trd.2019.07.005, 2019.Mijić, Z., Stojić, A., Perišić, M., Rajšić, S., and Tasić, M.: Receptor modeling studies for the characterization of PM10 pollution sources in Belgrade, CI&CEQ, 18, 623–634, 10.2298/CICEQ120104108M, 2012.Mirakovski, D., Boev, B., Boev, I., Hadzi-Nikolova, M., Reka, A., and Sijakova-Ivanova, T.: Urban air pollution in Skopje aglomeration–trafic vs background case, Contrib. Sect. Nat. Math. Biotech. Sci., MASA, 40, 41–48, e-ISNN 1857-9949, https://eprints.ugd.edu.mk/26594/1/Prilozi-41-12020-1.pdf (last access: 8 August 2026), 2020.Moreno, T., Querol, X., Castillo, S., Alastuey, A., Cuevas, E., Herrmann, L., Mounkaila, M., Elvira, J., and Gibbons, W.: Geochemical variations in aeolian mineral particles from the Sahara–Sahel Dust Corridor, Chemosphere, 65, 261–270, 10.1016/j.chemosphere.2006.02.052, 2006.Newell, K., Kartsonaki, C., Lam, K. B. H., and Kurmi, O. P.: Cardiorespiratory health effects of particulate ambient air pollution exposure in low-income and middle-income countries: a systematic review and meta-analysis, Lancet Planet. Health, 1, e368–e380, 10.1016/S2542-5196(17)30166-3, 2017.Nicolás, J., Chiari, M., Crespo, J., Orellana, I. G., Lucarelli, F., Nava, S., Pastor, C., and Yubero, E.: Quantification of Saharan and local dust impact in an arid Mediterranean area by the positive matrix factorization (PMF) technique, Atmos. Environ., 42, 8872–8882, 10.1016/j.atmosenv.2008.09.018, 2008.Norris, G., Duvall, R., Brown, S., and Bai, S.: EPA positive matrix factorization (PMF) 5.0 fundamentals and user guide, US Environmental Protection Agency, Washington, DC, https://www.epa.gov/sites/default/files/2015-02/documents/pmf_5.0_user_guide.pdf (last access: 8 August 2026), 2014.Ooki, A., Uematsu, M., Miura, K., and Nakae, S.: Sources of sodium in atmospheric fine particles, Atmos. Environ., 36, 4367–4374, 10.1016/S1352-2310(02)00341-2, 2002.Paatero, P. and Hopke, P. K.: Rotational tools for factor analytic models, J. Chemom., 23, 92–100, 10.1002/cem.1197, 2009.Paatero, P. and Tapper, U.: Positive matrix factorization: A non-negative factor model with optimal utilization of error estimates of data values, Environmetrics, 5, 111–126, 10.1002/env.3170050203, 1994.Pandolfi, M., Alastuey, A., Pérez, N., Reche, C., Castro, I., Shatalov, V., and Querol, X.: Trends analysis of PM source contributions and chemical tracers in NE Spain during 2004–2014: a multi-exponential approach, Atmos. Chem. Phys., 16, 11787–11805, 10.5194/acp-16-11787-2016, 2016.Pandolfi, M., Mooibroek, D., Hopke, P., van Pinxteren, D., Querol, X., Herrmann, H., Alastuey, A., Favez, O., Hüglin, C., Perdrix, E., Riffault, V., Sauvage, S., van der Swaluw, E., Tarasova, O., and Colette, A.: Long-range and local air pollution: what can we learn from chemical speciation of particulate matter at paired sites?, Atmos. Chem. Phys., 20, 409–429, 10.5194/acp-20-409-2020, 2020.Pereira, G. M., Yoshiaki Kamigauti, L., Pereira, R. F., Monteiro dos Santos, D., da Silva Santos, T., Martins, J. V., Alves, C., Gonçalves, C., Casotti Rienda, I., Kováts, N., Nogueira, T., Rizzo, L., Artaxo, P., Maura de Miranda, R., Yamasoe, M. A., Dias de Freitas, E., de Castro Vasconcellos, P., and de Fatima Andrade, M.: Source apportionment and ecotoxicity of PM2.5 pollution events in a major Southern Hemisphere megacity: influence of a biofuel-impacted fleet and biomass burning, Atmos. Chem. Phys., 25, 4587–4616, 10.5194/acp-25-4587-2025, 2025.Perrone, M. G., Vratolis, S., Georgieva, E., Török, S., Šega, K., Veleva, B., Osan, J., Bešlić, I., Kertész, Z., Pernigotti, D., Eleftheriadis, K., and Belis, C. A.: Sources and geographic origin of particulate matter in urban areas of the Danube macro-region: The cases of Zagreb (Croatia), Budapest (Hungary) and Sofia (Bulgaria), Sci. Total Environ., 619, 1515–1529, 10.1016/j.scitotenv.2017.11.092, 2018.Pio, C., Cerqueira, M., Harrison, R. M., Nunes, T., Mirante, F., Alves, C., Oliveira, C., Sanchez de la Campa, A., Artinano, B., and Matos, M.: OC / EC ratio observations in Europe: Re-thinking the approach for apportionment between primary and secondary organic carbon, Atmos. Environ., 45, 6121–6132, 10.1016/j.atmosenv.2011.08.045, 2011.Pisso, I., Sollum, E., Grythe, H., Kristiansen, N. I., Cassiani, M., Eckhardt, S., Arnold, D., Morton, D., Thompson, R. L., Groot Zwaaftink, C. D., Evangeliou, N., Sodemann, H., Haimberger, L., Henne, S., Brunner, D., Burkhart, J. F., Fouilloux, A., Brioude, J., Philipp, A., Seibert, P., and Stohl, A.: The Lagrangian particle dispersion model FLEXPART version 10.4, Geosci. Model Dev., 12, 4955–4997, 10.5194/gmd-12-4955-2019, 2019.Platt, S. M., Davidović, M., Bartonova, A., Ćirović, Ž., Eckhardt, S., Evangeliou, N., Gundersen, H., Jovanović, M., Jovašević-Stojanović, M., Močnik, G., Petrović, B., Schneider, P., and Yttri, K. E.: Measurement report: Sources of carbonaceous aerosol in a South-East European metropolis, EGUsphere [preprint], 10.5194/egusphere-2026-1708, 2026.Polichetti, G., Cocco, S., Spinali, A., Trimarco, V., and Nunziata, A.: Effects of particulate matter (PM10, PM2.5 and PM1) on the cardiovascular system, Toxicology, 261, 1–8, 10.1016/j.tox.2009.04.035, 2009.Polissar, A. V., Hopke, P., K., Paatero, P., Malm, W., C., and Sisler, J., F.: Atmospheric aerosol over Alaska: 2. Elemental composition and sources, J. Geophys. Res.-Atmos., 103, 19045–19057, 10.1029/98JD01212, 1998.Pope III, C. A., Burnett, R. T., Thun, M. J., Calle, E. E., Krewski, D., Ito, K., and Thurston, G.: Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution, JAMA, 287, 1132–1141, 10.1001/jama.287.9.1132, 2002.Querol, X., Alastuey, A., Rodriguez, S., Plana, F., Ruiz, C. R., Cots, N., Massagué, G., and Puig, O.: PM10 and PM2.5 source apportionment in the Barcelona Metropolitan Area, Catalonia, Spain, Atmos. Environ., 35, 6407–6419, 10.1016/S1352-2310(01)00361-2, 2001.Querol, X., Viana, M., Alastuey, A., Amato, F., Moreno, T., Castillo, S., Pey, J., de la Rosa, J., Sanchez de la Campa, A., Artinano, B., Salvador, P., Dos Santos, S. G., Fernandez-Patier, R., Moreno-Grau, S., Negral, L., Minguillon, M. C., Monfort, E., Gil, J. I., Ortega, L. A., Santamaria J. M., and Zabalza, J.: Source origin of trace elements in PM from regional background, urban and industrial sites of Spain, Atmos. Environ., 41, 7219–7231, 10.1016/j.atmosenv.2007.05.022, 2007.Rajšić, S., Mijić, Z., Tasić, M., Radenković, M., and Joksić, J.: Evaluation of the levels and sources of trace elements in urban particulate matter, Environ. Chem. Lett., 6, 95–100, 10.1007/s10311-007-0115-0, 2008.Rowell, R. M., Pettersen, R., and Tshabalala, M. A.: Handbook of wood chemistry and wood composites, CRC press, https://www.academia.edu/download/112691927/b12487-5.pdf (last access: 8 August 2026), 2012.Ruuskanen, J., Tuch, T., Ten Brink, H., Peters, A., Khlystov, A., Mirme, A., Kos, G. P. A., Brunekreef, B., Wichmann, H. E., Buzorius, G., Vallius, M., Kreyling, W. G., and Pekkanen, J.: Concentrations of ultrafine, fine and PM2.5 particles in three European cities, Atmos. Environ., 35, 3729–3738, 10.1016/S1352-2310(00)00373-3, 2001.Samaké, A., Jaffrezo, J.-L., Favez, O., Weber, S., Jacob, V., Canete, T., Albinet, A., Charron, A., Riffault, V., Perdrix, E., Waked, A., Golly, B., Salameh, D., Chevrier, F., Oliveira, D. M., Besombes, J.-L., Martins, J. M. F., Bonnaire, N., Conil, S., Guillaud, G., Mesbah, B., Rocq, B., Robic, P.-Y., Hulin, A., Le Meur, S., Descheemaecker, M., Chretien, E., Marchand, N., and Uzu, G.: Arabitol, mannitol, and glucose as tracers of primary biogenic organic aerosol: the influence of environmental factors on ambient air concentrations and spatial distribution over France, Atmos. Chem. Phys., 19, 11013–11030, 10.5194/acp-19-11013-2019, 2019. Seinfeld, J. H. and Pandis, S. N.: Atmospheric chemistry and physics: from air pollution to climate change, 3rd edn., John Wiley & Sons, ISBN 978-1-119-22117-3, 2016.Serbian Environmental Protection Agency: Annual Report on the State of Air Quality in the Republic of Serbia for 2023, Serbian Environmental Protection Agency, Ministry of Environmental Protection, Republic of Serbia, https://sepa.gov.rs/wp-content/uploads/2024/10/Vazduh2023.pdf (last access: 8 August 2026), 2024.Serbian Environmental Protection Agency: Annual Report on the State of Air Quality in the Republic of Serbia for 2024, Serbian Environmental Protection Agency, Ministry of Environmental Protection, Republic of Serbia, https://sepa.gov.rs/wp-content/uploads/2026/01/Vazduh2024.pdf (last access: 8 August 2026), 2026.Schmidl, C., Marr, I. L., Caseiro, A., Kotianová, P., Berner, A., Bauer, H., Kasper-Giebl, A., and Puxbaum, H.: Chemical characterisation of fine particle emissions from wood stove combustion of common woods growing in mid-European Alpine regions, Atmos. Environ., 42, 126–141, 10.1016/j.atmosenv.2007.09.028, 2008.Srivastava, D., Tomaz, S., Favez, O., Lanzafame, G. M., Golly, B., Besombes, J. L., Alleman, L. Y., Jaffrezo, J. L., Jacob, V., Perraudin, E., Villenave, E., and Albinet, A.: Speciation of organic fraction does matter for source apportionment. Part 1: A one-year campaign in Grenoble (France), Sci. Total Environ., 624, 1598–1611, 10.1016/j.scitotenv.2017.12.135, 2018.Stafoggia, M., Oftedal, B., Chen, J., Rodopoulou, S., Renzi, M., Atkinson, R. W., Bauwelinck, M., Klompmaker, J. O., Mehta, A., Vienneau, D., Andersen, Z. J., Bellander, T., Brandt, J., Cesaroni, G., De Hoogh, K., Fecht, D., Gulliver, J., Hertel, O., Hoffmann, B., Hvidtfeldt, U. A., Jockel, K. H., Jorgensen, J. T., Katsouyanni, K., Ketzel, M., Kristoffersen, D. T., Lager, A., Leander, K., Liu, S., Ljungman, P. L. S., Nagel, G., Pershagen, G., Peters, A., Raaschou-Nielsen, O., Rizzuto, D., Schramm, S., Schwarze, P. E., Severi, G., Sigsgaard, T., Strak, M., van der Schouw, Y. T., Verschuren, M., Weinmayr, G., Wolf, K., Zitt, E., Samoli, E., Forastiere, F., Brunekreef, B., Hoek, G., and Janssen, N. A. H.: Long-term exposure to low ambient air pollution concentrations and mortality among 28 million people: results from seven large European cohorts within the ELAPSE project, Lancet Planet. Health, 6, e9–e18, 10.1016/S2542-5196(21)00277-1, 2022.Stojić, A., Stojić, S. S., Reljin, I., Čabarkapa, M., Šoštarić, A., Perišić, M., and Mijić, Z.: Comprehensive analysis of PM10 in Belgrade urban area on the basis of long-term measurements, Environ. Sci. Pollut. Res., 23, 10722–10732, 10.1007/s11356-016-6266-4, 2016.Terzi, E., Argyropoulos, G., Bougatioti, A., Mihalopoulos, N., Nikolaou, K., and Samara, C.: Chemical composition and mass closure of ambient PM10 at urban sites, Atmos. Environ., 44, 2231–3329, 10.1016/j.atmosenv.2010.02.019, 2010.Tian, F., Qi, J., Wang, L., Yin, P., Qian, Z. M., Ruan, Z., Liu, J., Liu, Y., McMillin, S. E., Wang, C., Lin, H., and Zhou, M.: Differentiating the effects of ambient fine and coarse particles on mortality from cardiopulmonary diseases: A nationwide multicity study, Environ. Int., 145, 106096, 10.1016/j.envint.2020.106096, 2020.Todorović, M. N., Radenković, M. B., Onjia, A. E., and Ignjatović, L. M.: Characterization of PM2.5 sources in a Belgrade suburban area: a multi-scale receptor-oriented approach, Environ. Sci. Pollut. Res., 27, 41717–41730, 10.1007/s11356-020-10129-z, 2020.Tonon, T., Li, Y., and McQueen-Mason, S. Mannitol biosynthesis in algae: more widespread and diverse than previously thought, New Phytol., 213, 1573–1579, 10.1111/nph.14358, 2017.Turpin, B. J. and Lim, H. J.: Species contributions to PM2.5 mass concentrations: Revisiting common assumptions for estimating organic mass, Aerosol Sci. Technol., 35, 602–610, 10.1080/02786820119445, 2001.Tursun, K., Omarova, A., Ibragimova, O., P., Bukenov, B., Tursumbayeva, M., Mukhtarov, R., Radelyuk, I., Yenisoy, S., Karakas, D., Ergin, H., Karaca, F., and Baimatova, N.: Dominant sources of PM2.5 in Kazakhstan's urban cities: A PMF and HYSPLIT-based study for air quality management in Central Asia, Urban Clim., 64, 102706, 10.1016/j.uclim.2025.102706, 2025.van Drooge, B. L. and Grimalt, J. O.: Particle size-resolved source apportionment of primary and secondary organic tracer compounds at urban and rural locations in Spain, Atmos. Chem. Phys., 15, 7735–7752, 10.5194/acp-15-7735-2015, 2015.Waked, A., Favez, O., Alleman, L. Y., Piot, C., Petit, J.-E., Delaunay, T., Verlinden, E., Golly, B., Besombes, J.-L., Jaffrezo, J.-L., and Leoz-Garziandia, E.: Source apportionment of PM10 in a north-western Europe regional urban background site (Lens, France) using positive matrix factorization and including primary biogenic emissions, Atmos. Chem. Phys., 14, 3325–3346, 10.5194/acp-14-3325-2014, 2014.Weber, S., Salameh, D., Albinet, A., Alleman, L. Y., Waked, A., Besombes, J. L., Jacob, V., Guillaud, G., Meshbah, B., Rocq, B., Hulin, A., Dominik-Segue, M., Chretien, E., Jaffrezo, J. L., and Favez, O.: Comparison of PM10 sources profiles at 15 French sites using a harmonized constrained positive matrix factorization approach, Atmosphere, 10, 310, 10.3390/atmos10060310, 2019.World Health Organization: Ambient (outdoor) air pollution, https://www.who.int/news-room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health (last access: 4 December 2025), 2024.Yao, W., Zhao, Y., Chen, R., Wang, M., Song, W., and Yu, D.: Emissions of toxic substances from biomass burning: a review of methods and technical influencing factors, Process., 11, 853, 10.3390/pr11030853, 2023.Yttri, K.: Organic tracers at Beograd Ada Marina, EBAS [data set], 10.48597/AJ4A-EZ64 (last access: 8 August 2026), 2024.Yttri, K. E., Simpson, D., Stenström, K., Puxbaum, H., and Svendby, T.: Source apportionment of the carbonaceous aerosol in Norway – quantitative estimates based on 14C, thermal-optical and organic tracer analysis, Atmos. Chem. Phys., 11, 9375–9394, 10.5194/acp-11-9375-2011, 2011.Yttri, K. E., Canonaco, F., Eckhardt, S., Evangeliou, N., Fiebig, M., Gundersen, H., Hjellbrekke, A.-G., Lund Myhre, C., Platt, S. M., Prévôt, A. S. H., Simpson, D., Solberg, S., Surratt, J., Tørseth, K., Uggerud, H., Vadset, M., Wan, X., and Aas, W.: Trends, composition, and sources of carbonaceous aerosol at the Birkenes Observatory, northern Europe, 2001–2018, Atmos. Chem. Phys., 21, 7149–7170, 10.5194/acp-21-7149-2021, 2021.Yttri, K. E., Bäcklund, A., Conen, F., Eckhardt, S., Evangeliou, N., Fiebig, M., Kasper-Giebl, A., Gold, A., Gundersen, H., Myhre, C. L., Platt, S. M., Simpson, D., Surratt, J. D., Szidat, S., Rauber, M., Tørseth, K., Ytre-Eide, M. A., Zhang, Z., and Aas, W.: Composition and sources of carbonaceous aerosol in the European Arctic at Zeppelin Observatory, Svalbard (2017 to 2020), Atmos. Chem. Phys., 24, 2731–2758, 10.5194/acp-24-2731-2024, 2024.Zangrando, R., Barbaro, E., Kirchgeorg, T., Vecchiato, M., Scalabrin, E., Radaelli, M., Dordjevic, D., Barbante, C., and Gambaro, A.: Five primary sources of organic aerosols in the urban atmosphere of Belgrade (Serbia), Sci. Total Environ., 571, 1441–1453, 10.1016/j.scitotenv.2016.06.188, 2016.Zhai, J., Shao, S., Yang, X., Zeng, Y., Fu, T. M., Zhu, L., Shen, H., Ye, J., Wng, C., and Tao, S.: Chemically Resolved Respiratory Deposition of Ultrafine Particles Characterized by Number Concentration in the Urban Atmosphere, Environ. Sci. Technol., 58, 16507–16516, 10.1021/acs.est.4c03279, 2024. Zotter, P., Ciobanu, V. G., Zhang, Y. L., El-Haddad, I., Macchia, M., Daellenbach, K. R., Salazar, G. A., Huang, R.-J., Wacker, L., Hueglin, C., Piazzalunga, A., Fermo, P., Schwikowski, M., Baltensperger, U., Szidat, S., and Prévôt, A. S. H.: Radiocarbon analysis of elemental and organic carbon in Switzerland during winter-smog episodes from 2008 to 2012 – Part 1: Source apportionment and spatial variability, Atmos. Chem. Phys., 14, 13551–13570, 10.5194/acp-14-13551-2014, 2014.