The role of fine and submicron aerosol particles in urban air pollution in the context of meeting new air quality standards

Description of the studied area and measurement campaign

The measurements were carried out at the station belonging to the Institute of Environmental Engineering of the Polish Academy of Sciences (IEE PAS), which meets the requirements for urban background locations^21,22. The research area is situated within the Silesia Province (Southern Poland), specifically in the heart of the Upper Silesian Industrial District (USID) (Fig. 1)—one of the most densely populated and most industrialized parts of the country^26,27. The immediate surroundings of the station are²⁷: (i) to the north (~ 400 m): a national road no 88 with heavy traffic; (ii) to the east (~ 200–300 m): provincial road no 921, commercial and service facilities, residential buildings; (iii) to the south: blocks of flats and single-family housing (~ 200–300 m) and at a further distance (~ 1 km) the city center with residential and commercial buildings; (iv) to the west (~ 200 m): housing estates and allotments.

The measurement campaign using the FIDAS spectrometer (see Sect. 3.2) was launched at the end of May 2022, in connection with the implementation of the statutory tasks of IEE PAS related to the assessment of the role of fine and ultrafine particles in air pollution in urban and non-urban areas. The research has been conducted continuously to this day, except for a longer break from April 20 to June 8, 2024, related to equipment repairs and calibration. For the purposes of this study, data from June 2022 to May 2025—covering three full calendar years—were analyzed. In order to assess the temporal variability of measured air pollutant concentrations, the entire measurement campaign was divided into shorter periods, i.e.: (i) years, where Year 1 covers the period Jun 2022 – May 2023, Year 2: Jun 2023 – May 2024, Year 3: Jun 2024 – May 2025; (ii) specific periods related to changes in the intensity of anthropogenic emission sources (especially fuel combustion in households), i.e., heating season (H) (Jan–Mar, Oct–Dec) and non-heating season (NH) (Apr–Sep); (iii) seasons, i.e., summer (Jun–Augt), autumn (Sep–Nov), winter (Dec–Feb), spring (Mar–May); (iv) individual months of the annual period. The adopted terminology is continued later in the article to maintain the clarity of the work.

Measuring equipment

The FIDAS^®200 S fine dust aerosol spectrometer (Palas GmbH) was developed for the monitoring of air pollution in accordance with the requirements of the Air Quality Directives^22,23 and national regulations of EU countries in this field. The device continuously measures fine aerosol particles in the ambient air and calculates PM₁₀ and PM_2.5 concentrations [in µg m^{− 3}], which are subject to measurement in accordance with applicable law. Concurrently, it calculates and records the concentrations of other PM fractions [in µg m^{− 3}]—PM₁, PM₄, and total suspended particulates (TSP), as well as particle number concentrations (C_N, in #cm³) and particle size distribution. However, the latter was not the subject of this article.

FIDAS^®200 S is based on the optical method, specifically light scattering on individual aerosol particles. In fact, the single particles move through an optically differentiated measurement volume that is homogeneously illuminated with a polychromatic LED light source. Each particle generates a scattered light impulse that is detected at an angle of 85° to 95° degrees. The optical aerosol sensor, which is the central part of the device, measures the intensity of this scattered light with the impulse’s intensity corresponding to the particle’s size and the number of impulses indicating the particle count. These results are then used to calculate the particle size distributions and mass concentrations of specific PM fractions. All measurement data is collected locally in FIDAS’s internal memory. Additionally, the system is connected to a datalogger manufactured by DAC System, which allows direct access to the data stream, remote viewing, and archiving on an external server.

Due to its high time resolution (1–24 h), low maintenance requirements, and relatively low overall operating costs, the apparatus has been successfully used worldwide for air quality monitoring^28–30. The measurement range declared by the manufacturer is 0–20,000 particles per cm³ (particle number), 0.18–18 μm (particle size), and 0–10,000 µg m^{− 3} (particle mass). An additional advantage of the device is the ability to simultaneously measure basic meteorological parameters (temperature, relative humidity, atmospheric pressure, wind speed and direction, intensity and type of precipitation), thanks to a meteorological station (LUFFT WS600) connected to the spectrometer. All measurements are performed in real time, which enables the analysis of the temporal variability of the measured pollutant concentrations—both in the short term (daily fluctuations) and long term (seasonal variations)—and, in combination with meteorological measurements, the identification of potential inflow directions (emission sources).

Data processing

All data collected over the 3-year measurement period was stored in the internal IEE PAS database. The MSExcel (Redmond, Washington, DC, USA) and Statistica 13.3 (Tibco Software Inc, San Ramon, CA, USA) software packages were used for statistical analysis. First, the time coverage of datasets was assessed, which determines the percentage of valid data for the considered averaging period. For this purpose, reference was made to the provisions of Directive 2024/2881, exactly Annex V, which specifies the required data quality objectives. For PM₁₀ and PM_2.5 measurements, the minimum time coverage is 85% (annual, 1-hour and 24-hour means), and for UFPs and their particle size distribution—80% (only annual means). The minimum percentage of valid data when aggregating them to calculate statistical parameters is 75%, both for 1-hour and 24-hour averages.

In the case of measurements performed with the FIDAS spectrometer (CN, PM₁, PM_2.5, PM₄, PM₁₀, TSP, meteorological parameters), the database was a set of raw data recorded with a time resolution of 1 min. On their basis, 1-hour average levels were calculated, which were then used to determine average daily level and then, monthly, seasonal, and annual levels, based on which tabular and graphical summaries were prepared in Sect. 3. Regarding the data on the concentrations of gaseous pollutants, obtained from the website of the Chief Inspectorate of Environmental Protection (CIEP) (see Sect. 2.4), the database included 1-hour average concentrations, which were subjected to the same processing as FIDAS results.

The authors of the article decided to additionally look at the periods of very high PM concentrations, also often called PM episodes or smog (Sect. 3.3), which are particularly critical from the point of view of air quality and short-term population exposure. Identifying episodes of elevated pollutant concentrations is not an easy task and is generally based on the coexistence of two factors^15,31: (i) sufficiently high pollutant concentrations (e.g. exceeding the permissible level); (ii) the occurrence of a set of meteorological factors responsible for these high concentrations, of which the most important are: high atmospheric pressure system, average daily air temperature < 5 °C, average daily wind speed < 3 m s^{− 1}, wind direction, average daily relative humidity > 85%, and no or very low atmospheric precipitation. The above-mentioned criteria are not rigid and the qualification of a day as a situation with an episode of high concentrations is usually made individually for a given area. In this study, episodes were defined based on literature reports and the authors’ previous experience^24,31—as cases with the daily PM_2.5 concentration exceeding 50 µg m^{− 3} for at least 2–3 consecutive days. The level of 50 µg m^{− 3} is also indicated in the new air quality directive as the information and alert threshold for fine particulate matter (for the PM₁₀ fraction both are 90 µg m^{− 3}).

Meteorological situation and the concentrations of gaseous pollutants

Numerous scientific studies have indicated that analysis of local weather conditions, such as air temperature, relative humidity, precipitation, wind speed and direction, is necessary for an in-depth understanding of the temporal and spatial variability of air pollution levels^12,18,25. General information about the meteorological situation during the measurement campaign is summarized in Table 1. The average outdoor air temperature during the 3-year period was 10.89 °C (10.47–11.43 °C in individual years) and was higher compared to the average annual temperature for the multi-year period 1991–2020 (8.8 °C), which is consistent with the general increasing temperature trend observed in Poland³². Temperature levels showed typical seasonal variations, with the highest values ​​in the summer (average: 20.35 °C) and the lowest in the winter (average: 2.36 °C). On the contrary, relative humidity showed greater fluctuations and lower levels in spring and summer (average: ~67–68%), while in autumn and winter it remained relatively constant and higher (average: 79–81%). Atmospheric pressure in the study area should be considered typical for Southern Poland³², with mean annual values ​​of ~ 1014–1018 hPa. Generally, during the measurement campaign, weak and mild winds dominated, according to the Beaufort scale, with mean levels in individual years of ~ 3–4 m s^{− 1}; however, the share of atmospheric calms on an annual basis was relatively low (~ 5–9%). Regardless of the measurement year, the study area was clearly dominated by southerly (S and SSE sectors) and northwestern (WNW sector) winds, with the former prevailing in the spring and summer periods, and the latter in the autumn and winter months.

This section can be considered as a reference point for a detailed analysis of episodes of elevated PM concentrations (see Sect. 3.3), which are special periods in many respects²⁴. In addition to meteorological parameters, Table 1 also presents the average concentrations of selected gaseous pollutants, based on data obtained from the general public database of the CIEP (https://powietrze.gios.gov.pl/pjp/rwms/12). They included the concentrations of sulfur dioxide (SO₂), nitrogen oxide (NO), nitrogen oxides (NO₂), nitrogen oxides (NO_X), and ozone (O₃), measured in parallel at the air quality monitoring station operated by the Regional Inspectorate for the Environmental Protection (RIEP) in Katowice, located next to the IEE PAS station (~ 5 m).

General overlook on PM concentrations

Studies conducted over a 3-year measurement period showed that both the particle number concentration (C_N) and the mass concentrations of individual fractions of particulate matter varied within wide limits (Fig. 2). Taking into account the entire measurement campaign, the hourly levels were in the range: 4.32–10777.50 #cm³ (C_N), 0.24–590.33 µg m^− 3 (PM₁), 0.41–620.03 µg m^− 3 (PM_2,5), 0.89–652.48 µg m^− 3 (PM₁₀), and 1.40–669.99 µg m^− 3 (TSP). The average daily concentrations were (Fig. 2): 57.85–4549.81 #cm³ (C_N), 2.02–239.09 µg m^− 3 (PM₁), 2.85–251.59 µg m^− 3 (PM_2,5), 4.44–263.31 µg m^− 3 (PM₁₀), and 4.88–270.23 µg m^− 3 (TSP). The average values ​​for the entire measurement period, calculated on the basis of 24-hour concentrations, were (Table 2): 582.13 ± 500.44 #cm³ (C_N), 22.27 ± 23.44 µg m^− 3 (PM₁), 24.49 ± 24.43 µg m^− 3 (PM_2,5), 30.77 ± 25.77 µg m^− 3 (PM₁₀), and 36.79 ± 27.62 µg m^− 3 (TSP). The concentrations of the measured variables in individual annual periods showed some differences, from relatively low levels in the 2nd year of measurements (AVE: C_N = 501.04 #cm³; PM₁ = 19.76 µg m^− 3; PM_2.5 = 22.05 µg m^− 3; PM₁₀ = 28.18 µg m^− 3; TSP = 33.78 µg m^− 3) to quite high in the 3rd year (AVE: C_N = 633.27 #cm³; PM₁ = 24.73 µg m^− 3; PM_2.5 = 27.08 µg m^− 3; PM₁₀ = 34.22 µg m^− 3; TSP = 41.30 µg m^− 3). This upward trend may raise some concern in the context of meeting the requirements of the new air quality directive, and identifying the possible causes of this situation requires additional research and analysis. It seems that the probable reason was the large smog situations that occurred in early December 2024 and in the first half of January 2025 (see discussion in Sect. 3.3), which covered the entire area of Southern Poland.

Referring to the latest report by the European Environment Agency²⁰, the concentrations of regulated PM fractions—PM₁₀ and PM_2.5—obtained in this study were higher than the values ​​recorded at most measurement stations in Europe, which means that the study area and the entire region of Southern Poland can still be considered one of the European hotspots for particulate air pollution. This applies in particular to fine PM—comparable or higher average annual PM_2.5 concentrations (i.e., > 20 µg m^{− 3}) were observed only at some sites in Italy, Turkey, and most of the Western Balkan countries. Similarly to PM₁₀, the main reason for this disparity is the increased use of solid fuels (hard coal, lignite, or wood) in central and eastern Europe, especially for heating households and in some industrial facilities or power plants²⁰. It is worth noting that PM₁₀ concentrations obtained from measurements conducted with the FIDAS^®200 S spectrometer were higher compared to those recorded at the CIEP station, located in close proximity to the IEE PAS site. However, based on the very high correlation between PM₁₀ concentrations from both sites (r = 0.96, α = 0.05), it should be concluded that the PM structure—i.e., the share of individual fractions in the total particulate mass (see Sect. 3.2)—was estimated reliably.

Literature data on PM₁ concentrations is much less available because no air quality standards have been established for this PM fraction, which means there is no monitoring obligation. Extensive information on PM₁ levels across European countries is provided in³³—PM₁ concentrations recorded at selected urban stations were (2013–2019): 4.4 µg m^{− 3} (Dublin, Denmark), 6.1 µg m^{− 3} (Helsinki, Finland), 8.7 µg m^{− 3} (Zurich, Switzerland), 9.0 µg m^{− 3} (Barcelona, Spain), 9.7 µg m^{− 3} (Paris, France), 11.5 µg m^{− 3} (Athens, Greece), 14.0 µg m^{− 3} (Lille, France), and 19.7 µg m^{− 3} (Bucharest, Romania). Concentrations obtained in this study were therefore significantly higher; only the level registered in Budapest can be considered similar. Furthermore, PM₁ levels noted in Zabrze during the period June 2022 – May 2025 were higher or comparable to those at most Polish measurement stations—e.g. Kraków (12 ± 5 µg m^{− 3}, spring 2016)¹⁴, Warsaw (11.07 and 17.41 µg m^{− 3} in summer 2014 and winter 2015, respectively)⁹, Szczawno-Zdrój (14.2 ± 3.4 µg m^{− 3} and 23.1 ± 8.8 µg m^{− 3} in summer and winter 2021, respectively)¹⁵ or Radom (26 ± 13 µg m^{− 3}, Feb 2020 – Apr 2021)¹¹.

Unlike the mass concentrations of PM₁ and PM_2.5, particle number concentrations obtained in this study were significantly lower than those recorded at many measurement sites worldwide. For example, Ridolfo et al. (2024)¹³ indicate that C_N recorded at urban background stations in Europe average 24,986 ± 12,171 #/cm³; in the US, C_N is generally lower (on average: 16,303 ± 5,374 #/cm³), which the authors attribute to the higher proportion of diesel vehicles in the former. Nevertheless, the same study emphasizes that such comparisons should be made with great caution due to the different measuring equipment and the different lower size detection limits of the devices used in the studies. In the case of the FIDAS^®200 S spectrometer, this value is 180 nm, so the finest aerosol particles, which may contribute significantly to the particle number concentration, are not taken into account.

In summary, the high aerosol concentrations recorded in this study are not surprising and should be explained by the location of the measurement site within the region of Southern Poland, where conditions of constant atmospheric stability prevail, and where there is a large accumulation of anthropogenic activities^25,26,31. Furthermore, the obtained results indicate that seasonal variability of PM concentrations should be taken into account when assessing population exposure to particulate air pollution. All the curves shown in Fig. 2 undergo similar temporal shifts throughout the period, or more precisely, they all follow the same annual cycle. The highest levels of C_N and individual PM fractions were recorded during the heating season, especially in the winter months, while the lowest levels were recorded in the spring and summer. The observed seasonal differences may be due to several factors, however, it is believed that high PM concentrations in autumn and winter, especially PM_2.5 and PM₁, are primarily due to municipal sources, whose activity increases at this time of year, while large point industrial emitters and traffic-related sources remain active throughout the year^2,34. The influence of meteorological conditions is also significant¹⁵. Higher PM concentrations during the heating season are associated with low air temperatures and lower wind speeds during this period, which in turn leads to a lower mixing layer height, thus limiting pollutant dispersion in the atmosphere. The decrease in PM_2.5 concentrations during the non-heating season results from conditions favoring pollutant dispersion, including higher air temperature and wind speed, which increase the mixing layer height, and more frequent and intense precipitation, which contributes to more efficient washing of atmospheric particles.

Finally, the identification of possible source regions of elevated PM levels was performed by the joint analysis of hourly PM concentrations and wind data (speed and direction). Radar plots were generated for all measured fractions and various averaging periods. However, because the concentration roses had a very similar pattern, Fig. 3 shows the distributions only for C_N, PM₁, and TSP (entire measurement period). In general, the distribution of PM₁, TSP, and C_N was quite directed, with the highest levels occurring primarily with the inflow of air masses from the eastern directions (maximum for the NE sector: C_N = 924.96 #/cm³; PM₁ = 36.51 µg m^− 3; TSP = 54.86 µg m^− 3). This suggests a significant role of road transport, commercial and service facilities, and residential development, as well as distant industrial emitters, allotment gardens, and the city center. The high impact of emitters located at a greater distance from the measurement site ( > ~ 2 km) on the particulate air pollution of the studied area is confirmed by the distribution of pollutant flows. In turn, very high concentrations of the measured substances recorded during periods considered as wind calms indicate a significant impact of the nearest local sources. This is typical for all background stations, which should evenly reflect the influence of different emission sources and can therefore serve as a reference point for measurements carried out at more specific locations (e.g. traffic, kerbside, or industrial sites).

Analysis of PM temporal variations in the context of meeting the requirements of air quality directives

Seasonal fluctuations of PM_2.5 and PM₁₀ concentrations and contributions were analyzed primarily to identify the most critical periods in terms of air quality. For these fractions, limit values ​​were established for average annual and daily concentrations, as well as the permissible frequency of exceedance of the daily PM₁₀ and PM_2.5 standard. The obtained results are important for assessing the health effects of air pollution and provide an essential scientific basis for future epidemiological studies.

Comparing the obtained results to air quality standards currently in force²², the limit value for the average annual PM₁₀ concentration (40 µg m^{− 3}) was not exceeded in any of the annual periods. However, the more stringent revised annual standard (20 µg m^{− 3}) was not met—PM₁₀ levels were on average ~ 149% (year 1), ~ 141% (year 2), and even ~ 171% (year 3) higher. Also noteworthy are the high concentrations of fine particulate matter compared to the established annual limit value (20 µg m^{− 3}, Directive 2008/50/EC). This disproportion is even more visible for the revised annual PM_2.5 standard (10 µg m^{− 3}) —the annual mean PM_2.5 concentrations recorded at the urban background station in Zabrze were higher by ~ 241%, ~ 221% and ~ 271%, in the 1st, 2nd and 3rd year of measurements, respectively.

For a large part of the measurement period, both the average daily concentrations of PM_2.5 and PM₁₀ remained above the limit values (Figs. 4 and 5). In total, during the entire measurement campaign, as many as 140 exceedances of the daily PM₁₀ standard currently in force (50 µg m^{− 3}), and 185 exceedances of the revised level (45 µg m^{− 3}) were recorded, which constituted ~ 13% and ~ 17% of the measurement period, respectively. The daily standard was not met in any of the annual periods, especially in relation to the new regulations²¹. In the 1st year of measurements, 63 days were recorded with PM₁₀ concentration above 45 µg m^{− 3} (~ 17% of measurement time) and 49 days with PM₁₀ concentration > 50 µg m^{− 3} (~ 13%); in the 2nd year, the number of exceedances was lower and amounted to 45 (~ 12%) and 35 (~ 10%), respectively. The 3rd year was the worst from the point of view of air quality—the annual PM₁₀ concentration was relatively high (34.21 µg m^{− 3}), as was the number of exceedances, amounting to 57 days (~ 21%) according to revised regulations.

The limit value for the average daily PM_2.5 concentration (25 µg m^{− 3}) was only introduced by Directive 2024/2881 and is higher than the value recommended in the WHO guidelines (15 µg m^{− 3})¹⁶. Analyzing the results in the context of compliance with this new standard, it was found that for over a quarter of the entire 3-year measurement period, average daily PM_2.5 concentrations exceeded the level of 20 µg m^{− 3}—a total of 334 exceedances were recorded. The situation varied between individual annual periods, but due to the high number of exceedances (> 18 days), the daily standard for PM_2.5 was not met in any year. Particularly unfavorable conditions—in terms of exposure of the residents to high concentrations of fine PM—occurred in the 1st and 3rd years of measurement, when 119 and 124 exceedances were recorded, respectively, representing ~ 34% of the measurement time. In the 2nd year of measurement, this number was lower (91 exceedances, ~ 25%), although still very high.

Analyzing the graphs in Figs. 5 and 6 and a certain pattern can be observed. Exceedances of daily standards for PM_2.5 and PM₁₀ generally did not occur during the summer, when concentrations of both fractions were relatively low, with the lowest monthly average levels recorded in July 2022 (PM_2.5 = 8.24 µg m^− 3; PM₁₀ = 14.69 µg m^− 3). With the arrival of the autumn, the situation deteriorated significantly—the number of exceedances of the daily PM_2.5 standard increased to 45, 19, and 30 days in the 1st, 2nd, and 3rd year of measurements, respectively; for PM₁₀, there were 22, 8 and 17 exceedances of the revised daily limit. The highest concentrations of PM_2.5 and PM₁₀, as well as other PM fractions, were observed in the winter months, with the highest monthly average levels observed in February 2025 (PM_2.5 = 61.86 µg m^− 3; PM₁₀ = 69.19 µg m^− 3). The number of exceedances of the daily PM_2.5 standard was as many as 54, 52, and 64 days, in the 1st, 2nd, and 3rd measurement year, respectively, which constituted ~ 14–18% of the time per year. In the case of the PM₁₀, the number of days with concentrations > 45 µg m^− 3 was 31 (year 1), 26 (year 2), and even 43 (year 3), with a share of ~ 7–12% annually. In the spring, the situation was slightly better than in autumn, with the number of exceedances of 20–23 days (PM_2.5) and 10–13 days (PM₁₀, 45 µg m^− 3).

A detailed analysis of the obtained results confirmed that seasonality has a significant influence on the concentrations of atmospheric aerosols in urban air, as found in numerous literature studies^9,11,17. Another issue that needs to be determined is: which fraction has the greatest impact on high PM concentrations and, consequently, will be the main contributor to exceeding air quality standards? Furthermore, will the shares of individual PM fractions in total particulate mass vary over the year, similarly to the concentrations? The almost identical seasonal fluctuations in the levels of PM₁, PM_2.5, PM₁₀, TSP as well as C_N (Sect. 3.1) were caused by comparable properties of the analyzed substances; moreover, all fractions are part of the TSP and will therefore be subject to the same temporal variations. The analysis of the variability of relative shares will provide greater scientific insight into the possible cause of high PM concentrations in the study area, constituting the important step in identifying probable emission sources.

The average shares of selected PM fractions were very similar in subsequent years, so Fig. 6 includes data from the entire measurement campaign. In general, PM in the air of the study area is characterized by the dominance of fine aerosol particles (AVE: PM_2.5/TSP = ~ 63%), especially submicron aerosol. The high shares of PM₁ in TSP, and therefore its significant role in air pollution, may indicate that exposure to ambient air has a very important impact on health, due to the fact that these particles are much more biologically active than the larger ones⁴. In fact, the PM₁/TSP ratio changed very clearly over the annual cycle, following changes in the mass concentration of PM₁ and other PM fractions. The highest values ​​of the PM₁/TSP ratio were observed in the winter season (AVE: 77.82%), which applies to all annual periods. Particularly high shares of PM₁ in TSP were recorded in December, with monthly average values ​​of 82.74% (Dec 2022), 80.43% (Dec 2023), and 78.53% (Dec 2024). The share of PM_2.5 was also high, averaging 87.74% (Dec 2022), 86.43% (Dec 2023), and 83.70% (Dec 2024).

Similar seasonal variations were also observed in the C_N, which was more than twice as high in the winter (AVE: 893.60 #/cm³) as in the summer (AVE: 325.12 #/cm³), with the highest average monthly level recorded in February 2025 (1373.44 #/cm³), followed by December 2024 (1074.57 #/cm³). Such differences in particle number levels (and high shares of PM₁ in TSP) have been indicated in literature studies^15,35, and are a consequence of more frequent inversion situations and enhanced particulate emissions during autumn-winter with respect to spring-summer. The observation site is influenced by many different types of anthropogenic sources, both those related to road traffic and domestic use of coal/wood/biomass burning for home heating. The greater atmospheric stability conditions and the reduction of the atmospheric boundary layer height, typical of the cold months, significantly impede the vertical dispersion of air pollutants that accumulate in the lower atmospheric layer, which in turn contributes to the increase of particulate matter mass and number concentrations³⁵.

The distinct seasonal variations in the share of coarser particles (d_ae > 2.5 μm) are indicative of their different origins and atmospheric processes. In general, the mean annual values ​​of the PM_2.5−10/TSP and PM_{10 − 18}/TSP ratios were very similar between years (~ 18–19%). In contrast to submicron particulate matter, the contributions of these fractions in the TSP mass were very low in the winter season, with monthly mean values ​​ranging from 6.94% (Dec 2022) to 13.32% (Feb 2024) (PM_2.5−10) and from 5.32% (Dec 2022) to 11.09% (Feb 2024) (PM_{10 − 18}). In turn, the highest shares of these fractions occurred in the summer season, together accounting on average for over half of the TSP mass (~ 56–57%). It seems that in the case of coarser particles, their concentrations are largely determined by sources such as mineral and soil dust or particles of biological origin, as the activity of these types of emission sources often increases in the warm season^30,33. The amount of precipitation and wind speed also play a major role, together influencing the dispersion and removal of pollutants²⁷. In fact, the fluctuations of the PM_{1 − 2.5}/TSP ratios were very different from those for PM₁/TSP, with monthly mean values ​​ranging from 4.40% (Feb 2025) to 8.52% (Oct 2023) and the average value over the entire measurement period at the level of 6.41%, which will indicate a similar origin.

Identification of high pollution periods

Continuing the discussion in Sect. 3.2 and based on the criteria presented in Sect. 2.3, periods of elevated pollutant concentrations were identified within this part of the study, paying particular attention to PM episodes. The decisive criterion was the average daily concentration of PM_2.5, as this fraction is among the regulated pollutants and its impact on health and the environment is considered more harmful compared to larger particles¹⁶.

During the entire 3-year measurement campaign, daily PM_2.5 concentrations below the limit value of 25 µg m^− 3 dominated, observed in a total of 714 days (~ 65% of the measurement time) (Table 3), in the following seasonal proportion: summer (262 days, ~ 95%), autumn (179 days, ~ 66%), spring (172 days, ~ 62%), winter (101 days, ~ 37%). During such periods, the air quality in the study area should be considered quite favorable, due to low concentrations not only of particulate but also gaseous pollutants (on average: NO₂ = 16.92 µg m^− 3; SO₂ = 6.55 µg m^− 3, among others). Moreover, the levels of substances covered by legal regulations were relatively low compared to the established limit values ​​for average annual concentrations; only in the case of PM_2.5 there was a slight violation of the new annual standard.

The number of days with PM_2.5 concentrations > 25 µg m^{− 3} was lower (334 days) (see Sect. 3.2). However, these periods determined the relatively high concentrations of fine PM in individual years, exceeding both the new and previous annual standards. High concentrations were recorded mainly in the heating season, in the following seasonal proportion: (i) 25 < PM_2.5 ≤ 50 µg m^{− 3}: winter (103, ~ 38%) < autumn (73, ~ 27%) < spring (42, ~ 15%) < summer (7, ~ 3%); (ii) 50 < PM_2.5 ≤ 100 µg m^{− 3}: winter (53, ~ 20%) < spring (21, ~ 8%) < autumn (20, ~ 7%); (iii) PM_2.5 > 100 µg m^{− 3}: winter (14, ~ 5%) < autumn (1, ~ 0.4%). As expected, as PM_2.5 concentrations increased, the concentrations of other fractions and particle number raised, also the levels of most gaseous substances. Of these, only ozone showed a decreasing trend, with average concentrations of 57.30 µg m^{− 3} (at PM_2.5 ≤ 25 µg m^{− 3}) and 15.67 µg m^{− 3} (at PM_2.5 > 100 µg m^{− 3}). This is a direct result of its atmospheric behavior—ozone is a typical secondary pollutant, and its maximum concentrations occur in the summer season, when high air temperatures and strong sunlight favor the occurrence of photochemical processes³⁶.

Of the 109 cases exceeding the level of 50 µg m^{− 3} (~ 10% of the measurement time), 84 days (~ 8%) were classified as “episode days”. A total of 26 such events were identified over the 3-year measurement campaign, which are highlighted in Fig. 7 with pale red shading and marked with the symbols “e” or “E,” respectively for “shorter” episodes (2 days) and those lasting 3 days or longer. The most days with episodes were recorded in the 1st measurement year (34 days, 9 events), followed by the 3rd year (31 days, 10 events), while the fewest days with episodes were recorded in the 2nd year—19 days and 7 events. Nevertheless, the concentration of gaseous pollutants (except ozone) also increased significantly during such events, including the regulated nitrogen dioxide (NO₂) and sulfur dioxide (SO₂). The mean NO₂ concentration during episodes was 33.80  µg  m^{− 3}, which constituted ~ 169% of the new annual standard for this substance (20 µg m^{− 3}, Directive 2024/2881). An even greater disproportion between episode and non-episode days was observed for nitric oxide (NO), with mean concentrations of 25.93 and 3.96 µg m^{− 3}, respectively, suggesting a significant impact of traffic emission sources²⁶. Interestingly, the SO₂ concentration during the episodes (AVE: 20.32 µg m^{− 3}) was only slightly higher compared to the annual standard (20 µg m^{− 3}, Directive 2024/2881). In general, SO₂ concentrations were relatively low throughout the measurement campaign, and scientific studies have already indicated the existence of a specific type of smog, called “Polish smog”, which is characterized by different chemical composition (low concentrations of SO₂) and different atmospheric conditions (high pressure, negative temperatures) compared to typical London smog²⁵. The decrease in air temperature and increase in atmospheric pressure during episodes were confirmed in this study, with mean values ​​of 1.60 °C and 1024.08 hPa, respectively, compared to 11.70 °C and 1015.76 hPa for non-episode days. The conditions for other meteorological parameters (see Sect. 2.3) also remained met. During periods with the highest PM concentrations, precipitation was absent or very low, as was wind speed, which averaged 2.63 m s^{− 1} compared to 3.78 m s^{− 1} for non-episode days. The increase in relative humidity was slightly less noticeable, although in individual episodes, daily mean levels exceeding 90% were recorded.

The most significant episode of high PM concentrations in the entire measurement campaign occurred at the end of December 2024 and lasted for 5 consecutive days (E18, Fig. 8). During this period, the highest concentrations of PM_2.5 were recorded (AVE: 169.51 µg m^− 3; MAX: 239.57 µg m^− 3), similarly to other particulate pollutants (AVE: C_N = 3435.38 #/cm³; PM₁ = 161.60 µg m^− 3; PM₁₀ = 177.05 µg m^− 3). They were also accompanied by very high concentrations of nitrogen oxides (NO_x) (AVE: 145.20 µg m^− 3), especially nitrogen oxide (NO) (AVE: 68.73 µg m^− 3), generally negative air temperatures (AVE: -0.54 °C) and high relative humidity, with average daily levels reaching up to 94%. Episode 13, recorded in the first half of January 2024, had only slightly lower average PM concentrations (C_N = 2955.56 #/cm³; PM₁ = 152.43 µg  m^− 3; PM_2.5 = 160.16 µg m^− 3; PM₁₀ = 167.28 µg m^− 3), while during this period the maximum daily levels of both number and mass concentrations of all suspended particulate matter fractions were noted. This period also stood out from the others due to very high concentrations of NO_x (AVE: 56.57 µg m^− 3) and SO₂ (AVE: 34.01 µg m^− 3) and the lowest air temperatures (AVE: -8.93 °C). The longest episodes were recorded in mid-December 2022 (E5), in the 2nd week of February 2023 (E7), and in the 3rd week of February 2025 (E22), with concentrations of particulate and gaseous pollutants remaining lower compared to the two episodes indicated above. However, due to the long duration (7 consecutive days), their role in high population exposure cannot be ignored. On the other hand, this also applies to “short-term” episodes—e.g. e2 (19-20.11.2022), e11 (30.11–01.12.2023), or e20 (03-04.02.2025) —with mean PM_2.5 concentration > 100 µg m^− 3 and C_N > 2000 #/cm³.

Figure 7 also highlights an additional period (pale yellow shading), recorded at the turn of March and April 2024. Because the PM_2.5 concentration was relatively low (AVE: 35.00 µg m^{− 3}) compared to high pollution days, it was not classified as a PM episode. The high TSP concentrations during this period (154.06–192.94 µg m^{− 3}) were determined by an unusually high share of coarser particles in total TSP (AVE: PM_2.5−10 = 44.94%; PM_{10 − 18} = 34.91%), which was not observed in any other period of the 3-year measurement campaign. Relatively high shares were also recorded for the PM_{1 − 2.5} fraction (AVE: 14.60%), while the share of the submicron PM was exceptionally low (AVE: 5.55%), similarly to the particle number concentration (C_N = 129.19 #/cm³). It seems that unlike PM episodes, conditioned by local factors (cold, calm weather, conducive to the accumulation of air pollutants), this event had an advection basis, likely associated with the influx of Saharan dust over the area of ​​central Europe. This assumption is confirmed by the work of³⁷, which indicated that from late March to early April 2024, most of mainland Europe was influenced by the intrusion of Saharan dust, experiencing an exceptionally high increase in air pollution.

To conclude the discussion in this section, it is worth taking another look at the dust structure in the study area—in terms of the percentage share of selected fractions in total particulate mass—but this time at different PM_2.5 concentration ranges (Fig. 8). During the clean days, with daily PM_2.5 concentrations > 10 µg m^{− 3}, coarser particles (d_ae > 2.5 μm) together accounted for more than half of the TSP mass, while the share of submicron particulate matter was relatively low (~ 40%), compared to the average value for the entire measurement period (~ 56%). Even a small increase in PM_2.5 concentrations changed this proportion, with fine PM (AVE: PM_2.5/TSP = ~ 58%) predominating over coarse PM and an average share of PM₁ in TSP of ~ 52%. With PM_2.5 levels exceeding the new daily standard, the dust structure changed dramatically. For the 25 < PM_2.5 ≤ 50 µg m^{− 3} range, the share of submicron particles increased to ~ 73%, while the share of coarser particles decreased two-fold. A further increase in PM concentrations was accompanied by an additional increase in the share of PM₁ in TSP, which averaged ~ 78% and ~ 80%, respectively, for the 50 < PM_2.5 ≤ 100 µg m^{− 3} range and during episodes. During periods of the highest air pollution with fine particulate matter (PM_2.5 > 100 µg m^{− 3}), the PM₁/TSP ratio reached its maximum, while the share of coarser particles in total aerosol mass did not exceed 10%. Therefore, the obtained results support the dominant anthropogenic origin of PM in the study area, in particular from traffic sources and so-called “low-level emission”, i.e., release of pollutants by emitters up to 40 m high, mainly from domestic furnaces (i.e., coal- and wood-fired boilers)^24,25. On the other hand, they indicate a strong need to include PM₁ measurements in routine air quality monitoring and the necessity to establish exposure limits for this fraction of particulate matter, especially in areas where PM_2.5 limit values ​​are frequently exceeded.

Competing interests

The authors declare no competing interests.