Particle Number Concentration: A Case Study for Air Quality Monitoring

: Particle matter is one of the criteria air pollutants which have the most considerable effect on human health in cities. Its legislation and regulation are mostly based on mass. We showed here that the total number of particles and the particle number concentrations in different size fractions seem to be efﬁcient quantities for air quality monitoring in urbanized areas. Particle number concentration ( N ) measurements were realized in Budapest, Hungary, for nine full measurements years between 2008 and 2021. The datasets were complemented by meteorological data and concentrations of criteria air pollutants. The annual medians of N were approximately 9 × 10 3 cm − 3 . Their time trends and diurnal variations were similar to other large continental European cities. The main sources of N are vehicle road trafﬁc and atmospheric new aerosol particle formation (NPF) and consecutive growth events. The latter process is usually regional, so it appears to be better assessible for contribution quantiﬁcation than mass concentration. It is demonstrated that the relative occurrence frequency of NPF was considerable, and its annual mean was around 20%. NPF events increased the contribution of ultraﬁne (UF < 100 nm) particles with respect to the regional particle numbers by 12% and 37% in the city center and in the near-city background, respectively. The pre-existing UF concentrations were doubled on the NPF event days.


Introduction
Air pollution is one of the most important factors affecting human health, the climate, and the environment. Around 91% of the world's population live in places with poor air quality. The ambient air pollution is estimated to account for 4.2 million premature deaths per year worldwide due mainly to stroke, lung cancer, heart disease, and acute and chronic respiratory diseases [1]. The sources of air pollution are multiple and complex. On a global scale, the major anthropogenic ambient sources include road vehicles, residential energy production for heating and cooking, power generation, industry, and agriculture [2]. The identification and characterization of the sources are crucial for understanding the effects of pollutants, as well as to develop suitable policies and technologies for moderating the air pollution, especially in cities [3].
Ambient air quality is ordinarily expressed by concentrations of certain key air pollutants and their health limit values. According to the World Health Organization (WHO), the key air pollutants include O 3 , NO 2 , SO 2 , CO, PM 2.5 mass, and PM 10 mass [1]. The latter two species express the particulate matter (PM) with aerodynamic diameters below 2.5 and 10 µm, respectively. As far as the health limits are concerned, there are global guidelines which also offer quantitative health-based recommendations for air quality management, namely, guidance on how to decrease the levels of these pollutants [1]. The European Environmental Agency (EEA) completed this list with NO x (= NO + NO 2 ), Pb, and benzene [4] and set their health limits. The U.S. Environmental Protection Agency (EPA) defined six outdoor criteria air pollutants, CO, Pb, ground-level O 3 , NO 2 , PM 2.5 or PM 10 masses, and SO 2 , and determined National Ambient Air Quality Standards (NAAQS) for them [5]. There are many further chemical species which are usually present in the ambient air and which can also cause harmful consequences for human health. The list of the key air pollutants can be further extended, e.g., by soot.
The regulatory issue of the PM is especially complicated since it is not a single chemical species, but it is a heterogeneous system which contains a complex mixture of more than a thousand of inorganic and organic compounds in their condensed phase dispersed in the air. In addition, as a colloidal system, it can be characterized by several different metrics that express important properties of particles. Hence, it cannot be expected that a single or a few metrics of the PM explain its comprehensive health effects. Particulate matter mass (in a certain size range), which is involved in the regulations, is one of the simplest quantities. It is usually associated with the health impacts of particles. Most epidemiological studies were based on mass as the dose metrics. The mass of atmospheric aerosols is, however, made of larger, i.e., coarse and fine particles. The mass contribution of smaller, e.g., of ultrafine (UF) particles (with an equivalent diameter <100 nm) is negligible. There are some PM types, atmospheric conditions, and specific health effects in which some other PM properties than the mass become important. These may include the number and surface area of particles. Many recent epidemiological and toxicological studies demonstrated that particle number concentrations, especially of UF particles, have a more considerable effect on human health than mass concentration [6][7][8][9][10][11][12][13][14][15]. UF particles, due to their size, can penetrate the respiratory system and even enter into the bloodstream and can cause inflammation and respiratory and cardiovascular diseases. This size fraction also represents an excess health risk relative to coarse and fine particles with the same or similar chemical composition [16,17]. It is worth mentioning that 70-80% of total particles in cities belong to the UF size range. It is, therefore, a plausible intention and requirement to extend the list of the key air pollutants by further aerosol metrics such as particle number concentration.
There have been mitigation policies and control regulations to reduce the emission of particle numbers as part of an overall air-quality improvement strategy since the 1990s. The legislation in the EU, including Hungary, focus, e.g., on particle emissions from diesel engines [18]. There were some important changes in the car emissions, which included the introduction of Euro 5 and 6 regulations for light-duty vehicles in January 2011 and Euro VI regulations for heavy-duty vehicles in September 2015 (the number of emitted particles with diameters >23 nm should be <6 × 10 11 km -1 ). The concentration of sulfur in diesel fuel for on-road transport was decreased in several phases to <10 ppm in January 2009 [19]. Sulfur content in fuels for mobile non-road diesel vehicles-including mobile machinery, agricultural and forestry tractors, inland waterway vessels, and recreational crafts-was limited to a level of 1000 ppm in 2008 and at 10 ppm in 2011. Dangerous fuel types for domestic heating are also listed, their emission factors are determined, and the accumulated information is disseminated among potential users. As far as secondary particles are concerned, it is not straightforward to reduce their concentration levels because the effects of gaseous and aerosol species are complex due to their nonlinear relationships and feedbacks in their related processes.
Total particle number concentrations are easily measured for monitoring purposes by condensation particle counters (CPCs), whereas particle number size distributions can be determined by online mobility particle size spectrometers. The latter systems possess the advantage that concentrations for different size fractions can be derived from the measured data. This is important, since different size fractions are related to different source types, atmospheric properties, and processes.
There are two major source types of particle numbers in the atmosphere: new particle formation and growth (NPF) events and high-temperature emissions. The former process is the dominant source in the global troposphere [20][21][22][23]. In cities and urbanized areas, the particle number concentrations are strongly affected by high-temperature emission sources from different sectors such as industrial processes, domestic installations, residential heating and cooking, vehicular road traffic, and power production [24,25]. In large cities, primary particles prevail over secondary particles [26][27][28]. Vehicles emit primary aerosol particles but also contribute to secondary aerosol particle formation by emitting their precursors [3]. New particle formation and growth events have been proven to be also be common in large cities, and they can have large particle number contributions on nucleation days [29]. Local meteorological conditions and long-range transport of air masses can play substantial roles in the concentrations actually realized [30]. It seems to be relevant and useful to investigate and overview the properties and behavior of particle number concentrations in longer data sets from the air quality aspects as well.
Atmospheric particle number concentrations in various size fractions and meteorological data for nine full measurement years are available for Budapest. Budapest is the capital of Hungary, which is located in the Carpathian Basin in Central Europe. It is the biggest and the most inhabited city of the country with around 525 km 2 of land area and has a population of 1.72 million inhabitants. The number of passenger cars registered in Budapest (596 × 10 3 in 2008 and 691 × 10 3 in 2020) increased slowly, while the share of the diesel-powered passenger cars enhanced somewhat more from approximately 20% in 2008 to 36% in 2020 [31]. The number of buses (ca. 4000) registered in Budapest and the share of the diesel-power buses (98%) on the national bus fleet remained constant.
The major objectives of this study are to overview the properties and time trends in particle number concentrations, to investigate their main sources, and to determine and discuss the contribution of NPF events to particle number concentrations. Further goals are to discuss the trends in nucleation source intensity and to interpret its consequences for the urban air quality.

Experimental Part
The measurements were performed in two different urban sites in Budapest. Most measurements were realized at the Budapest platform for Aerosol Research and Training (BpART) Laboratory (47 • 28 29.9 N, 19 • 3 44.6 E; 115 m above mean sea level, m.s.l.) of the Eötvös Loránd University. The location represents a well-mixed, average atmospheric environment for the city center due to its geographical and meteorological conditions [32]. Therefore, it can be regarded as an urban background site. The main local emissions are diffuse urban traffic exhaust, residential and household emissions, industrial sources, and some off-road transport [33]. The long-range transport of air masses can also play an important role. The other location was situated at the northwestern border of Budapest in a wooded area of the Konkoly Astronomical Observatory (47 •  The measurements in year Y2 were performed in the near-city background, while they were realized in the city center in the other years. As the time base of the data, local time (LT = UTC + 1 or daylight saving time, UTC + 2) was chosen because it has been observed that the daily activity of inhabitants significantly influences many atmospheric processes in the urbanized areas [34][35][36].
The particle number size distributions were determined by a laboratory-made flowswitching-type differential mobility particle sizer (DMPS) [37,38]. The system operates in the electrical mobility diameter range from 6 to 1000 nm in the dry state (with a relative humidity (RH) < 30%) of particles in 30 channels with a time resolution of 8 min [32,37]. Its main components include a Ni-60 radioactive bipolar charger, a Nafion semi-permeable membrane monotube dryer, a 28-cm-long Vienna-type differential mobility analyzer, and a butanol-based CPC (TSI, model 3775, USA, Shoreview, MN). The instrument was operated in two sets of flows: In the high flow mode, the aerosol flow rate was 2.0 L min −1 , and in the low mode, it was 0.31 L min −1 , while the sheath air flow rates were 10 times larger than the sample flows. The measurements were performed in a continuously way according to the international technical standards [39,40].
The CPC instrument (TSI, model 3752, USA, Shoreview, MN) operated with an aerosol inlet flow of 1.5 L min −1 and measured concentrations of particles with diameters above 4 nm using butanol as a working fluid. Mean particle number concentration data were extracted from its database with a time resolution of 1 min. The data were utilized for quality control of the integrated DMPS data.
The concentrations of SO 2 ; NO, NO x , and NO 2 ; O 3 ; CO; and PM 10 mass and PM 2.5 mass were measured by UV fluorescence (Ysselbach 43C, Budapest, Hungary), chemiluminescence (Thermo 42C, Waltham, MA, USA), UV absorption (Ysselbach 49C, Budapest, Hungary), IR absorption (Thermo 48i, Waltham, MA, USA), and beta-ray attenuation (two Environment MP101M instrument with PM 10 and PM 2.5 inlets) methods, respectively, with a time resolution of 1 h. The experimental data were acquired from the closest measurement stations of the National Air Quality Network in Budapest, which is located 4.5 km from the city center site and 6.9 km from the regional background site in the upwind prevailing wind direction [41].
Most meteorological measurements for the city center took place on-site. Air temperature (T), RH, wind speed (WS), and wind direction (WD) data were obtained by standardized meteorological methods (HD52.3D17, Delta OHM, Padova, Italy, and SMP3 pyranometer, Kipp and Zonen, Delft, The Netherlands, respectively) with a time resolution of 1 min (except for Y1, when it was 1 h). The meteorological data for the near-city background were measured using a mobile meteorological station installed at the measurement location at a height of ca. 2 m from the ground and the time resolution was 10 min. The data coverage for particle number concentrations, meteorological data, and criteria air pollutants were 94, >90, and >85%, respectively, over the whole interval.

Data Treatment
The measured DMPS data were inverted into discrete size distributions that were utilized to calculate particle number concentrations in the diameter ranges 6-25 nm (N 6-25 ), 25-100 nm (N 25-100 ), 6-100 nm (N 6-100 ), 100-1000 nm (N 100-1000 ), and 6-1000 nm (N 6-1000 ). These size intervals were selected to represent various important particle source types. The concentrations N 6-25 are associated with NPF events [20,26,42], N 25-100 are mostly emitted by incomplete combustion (such as vehicle road traffic or household heating) in urbanized areas or generated by condensational growth of new particles, and N 100-1000 mostly represents physically and chemically aged, thus regional particles [34,36]. The concentration N 6-100 (of UF particles) is of special interest since it is often related to excess health effects, while the N 6-1000 represents the total particles. The DMPS data were also utilized to generate daily particle number size distribution surface plots for identification and classification of the NPF events [43][44][45]. The following classes were defined: event days, non-event days, undefined days, and missing days. The relative occurrence frequency of NPF events (f NPF ) was determined individually for each month and year as the ratio of the number of event days to the total number of relevant days.
The importance and contribution of particles generated by NPF events were assessed by nucleation strength factors (NSFs) [34]. There are two factors, which are defined as follows: (1) It was implicitly assumed that the production of larger particles (>100 nm) was much smaller than the concentration of UF particles. This is typical for cities, and it can be proved by considering the contributions of UF particles to the total particle numbers [45,46]. The NSF NUC is determined exclusively for nucleation days, while the NSF GEN is derived for all available days. The former property represents the concentration increment from NPF on an ordinary nucleation day with respect to N 100-1000 , while the latter quantity expresses the overall increment in particles in general, thus on an average day [26]. Both NSFs were calculated separately for each month, each measurement year, and the whole measurement interval. Their diurnal variations were also derived.

Atmospheric Concentrations and Their Ratios
Basic descriptive statistics of the particle number concentrations in different relevant size fractions are summarized in Table 1. They demonstrate that the minimum concentrations in a selected size range were usually similar to each other over the years, while the other concentrations tended to be distinctly larger in the city center (Y1, Y3-Y9) than in the near-city background (Y2). This is due to the larger anthropogenic sources of particles in the city than in its surroundings. The vehicle road traffic represents the major contribution, and the traffic intensity is larger in the city center with respect to the near-city background. From this aspect, particle number concentration is indeed a valuable and useful quantity to express several important urban anthropogenic activities. The ranges of the concentrations were rather large; they could be expressed by a factor of 300. This is a specific feature of particle number concentrations, and it is mainly related to the dynamic character of their sources, atmospheric processes, and relatively short residence time of smaller (UF) particles. It is seen that there was a decreasing trend in the N 6-1000 and N 6-100 for the city center from 2008 over the years, and this was likely interrupted in 2016. The change in the N 100-1000 appeared to be more modest, which can be explained by the baseline character of this size range. The concentrations are in line with or similar to those in other large European cities [11,[47][48][49][50][51][52][53].
The N 6-100 represents the major contribution to the total particle numbers both in the city center and the near-city background [41]. Their mean values and SDs were (78 ± 10)% and (67 ± 16)%, respectively. In the city center, this is mostly due to the vehicle road traffic and other high-temperature emission sources. In urban areas, N 25-100 is also mainly composed of particles from high-temperature emission sources, which explains its larger temporal variability. The mean contribution and SDs of the chemically aged (regionally representative) particles to the total numbers were (22 ± 10)% and (33 ± 14)%, respectively.

Time Series
The annual time series of particle number concentrations in different size ranges during all measurement years are shown in Figure 1. We can see that there was no considerable monthly or seasonal variability of concentrations during a year. This is different from PM mass concentrations. It can be explained by relatively constant sources, atmospheric processes, and decrease in particle numbers over a year. The concentrations N 6-25 , N 25-100 , and N 6-1000 showed larger temporal variability, while the changes in N 100-1000 were more modest. This can be explained by the different major sources of size-fractionated particles, as discussed in Section 3.1. Table 1. Ranges, medians, and means with standard deviations (SDs) of particle number concentrations N 6-25 , N 25-100 , N 6-100 , N 100-1000 , and N 6-1000 (all in 10 3 cm −3 ) in the city center and near-city background for the measurement years Y1 to Y9.  The diurnal distribution of the air pollutants and particle number concentrations in different size ranges are presented in Figure 2. We can see that the concentrations of SO 2 and PM 2.5 mass did not change substantially during the day. They seem to have little relation with the vehicle road traffic. This is because the sources of the fine particles in Budapest are mostly related to non-vehicular processes [33]. In the case of the diurnal variation in PM 10 mass, we can see that it changed slowly and modestly during the day with tendencies for lower concentration overnight and higher levels during daylight. This can be related to their major sources such as the resuspension of urban dust, emissions from material wear from moving parts of vehicles, and the ageing of exhausted particles from vehicles and their relatively large atmospheric residence time [54]. At the same time, we can see that NO, NO 2 , and CO concentrations show two peaks corresponding to the typical behavior of traffic [55]. These pollutants are mainly related to vehicle road traffic. The peaks appeared around the typical rush hours in Budapest [55]. This property can be identified most evidently in the case of N 25-100 . This is explained by the fact that the main source of this size range (25-100 nm) in urban environments is incomplete combustion [36], thus vehicle traffic. On the diurnal pattern of N 6-25 , three peaks were identified; the first and the last can be associated with road traffic, while the peak around noon is linked to NPF events. The diurnal variation of N 100-1000 was more constant, and it only showed modest variability, as expected. All these features are reflected back on the diurnal variation of the total particle number. The diurnal series obviously showed associations among particle number concentrations and vehicle traffic. Furthermore, it was estimated that ca. 70% of total particle numbers in cities are generated by emissions [24]. This can, however, sensitively depend on and change with the local and regional atmospheric properties and conditions. Therefore, its contribution or importance are challenging to estimate. Instead of this, we assessed the contributions from NPF events in the present study since this process is related to a larger region at least in the study area, i.e., in the Carpathian Basin [38].

New Aerosol Particle Formation and Growth Events
The total number of NPF events in 9 years was 663. Its annual mean and SD were 74 ± 17. They resulted in an overall mean relative occurrence frequency and SD of (21 ± 4)%. It means that the phenomenon occurred with a considerable rate; there was an NPF event every fifth day on a yearly scale. This also suggests that the NPF events are an important source of particles even in cities. The numbers of nucleation days for the different months in each year are summarized in Table 2. The distributions of the monthly mean counts exhibited obvious differences, while the annual total counts were similar to each other in various aspects, except for year Y5, which also showed the smallest annual count. We could not find plausible explanation for this extremely small value. The other differences can likely be explained by substantial changes in multifactorial conditions and by the complex interplay among the influential environmental variables over a year and by inter-annual differences in chemical, aerosol and meteorological properties and in biogenic cycling [56][57][58]. The number of NPF event days in the near-city background location (Y2) was the highest among them. The realization of NPF events depends on the competition between the sources and sinks of condensing vapors expressed by their ratio [58]. They can still create favorable conditions for NPF occurrence for smaller source intensities if the condensation sink (CS) is even lower. This is expected to be the typical case for the near-city background site.
The mean f NPF values were calculated separately for days of the week and for workdays and holidays in the city center for 8 years (Figure 3). It can be seen that the values for the holidays were significantly larger than for the workdays or for the overall mean. On weekends, especially on Sundays, some anthropogenic sources are substantially reduced. For instance, the road traffic on weekends is decreased by approximately 30% with respect to workdays [55]. This results in smaller CS on these days, which appears to be favorable for the NPF event occurrence. As far as the workdays are concerned, they seem to fluctuate around their mean value. Mondays seem to exhibit somewhat lower frequency, which can likely be explained by usually larger traffic intensities on this day than on the other workdays. These results demonstrate that anthropogenic activities, in particular, vehicle road traffic, do affect the urban NPF phenomenon.

Contributions of NPF Events
The mean NSFs calculated separately for the city center and near-city background are summarized in Table 3. New particle formation (represented by the NSF NUC ) increased the particle number concentrations in the city center by a factor of 1.7 and by ca. 2 times in the near-city background. The importance of NPF on a longer time interval was demonstrated by the mean NSF GEN values. In the city center, 12% of UF particles were generated by NPF as a single source, while it produced 34% of UF particles in the near-city background. The relatively large SDs point to the changing intensity of NPF events. In addition, it is also seen that both the NSF NUC and NSF GEN values were systematically larger for the background than for the center. This is mainly due to the higher particle emissions in the city center, which is caused by anthropogenic activities. Relative occurrence frequency of new aerosol particle formation in the city center over 8 measurement years separately for days of the week and for workdays and holidays. The solid horizontal line indicates the overall mean, and the yellow band shows its standard deviation. The actual counts of the new particle formation event days are written above the columns. The annual mean NSFs for the city center are presented in Figure 4. The NSF GEN seems to be constant during the whole measurement interval. It is worth mentioning that the occurrence frequency did not change significantly during the investigated years (except for Y5, see Section 3.2). However, the annual mean NSF NUC values appear to indicate a slightly increasing tendency with a slope and SD of (2.6 ± 2.5)% annually. This may be just a fluctuation, or this may suggest that either the contribution of NPF events became larger or that the general level of N 100-1000 decreased during the years. The latter tendency cannot be confirmed in the corresponding concentration data from Table 1, and moreover, the possible changes in the annual mean NSF NUC and N 100-1000 are not in line or coherent with each other. This hints to possible increasing importance of NPF events with respect to emission sources in Budapest. The hypothesis should be further investigated by independent evaluation methods and on longer data sets.
The monthly mean NSF GEN values changed modestly on an annual scale. Their distribution basically followed the shape of the NPF occurrence frequency, which was shown in [58]. This can be explained by the fact that the monthly NPF contributions on a general day are expected to be larger if the NPF frequency is higher. It is more exciting to investigate the variations in the monthly mean NSF NUC , which are displayed in Figure 5. There seems to be a systematic tendency for lower values (down to 1.1) in summer and larger values (up to 2) in late autumn and early winter.  This is, however, a consequence of the seasonal trend in the N 100-1000 level. Its diurnal variations for the NPF days and non-NPF days in the non-winter seasons are presented in Figure 6a. The differences between the curves were negligible. In winter, the shapes were also similar to each other, while their magnitudes were considerably different (Figure 6b). The concentrations were substantially smaller on the NPF days than on the non-NPF days. In winter, the GRad and biogenic precursor gases in the air were decreased [58], and therefore, the source strength of the condensing vapors is expected to also be lower. Consequently, NPF events in this season occur only or preferably if the sink term (related mainly to the existing regional aerosol particles, thus to N 100-1000 ) is even smaller. This explains the difference between the diurnal curves for the NPF and non-NPF days, and the elevated NSF NUC values in winter (see Equation (1)). The NPF events rather took place on those winter days when the particle number concentrations were relatively smaller, and, therefore, the NPF increased the existing low concentration levels by a larger factor. The mean diurnal variations in the NSF NUC and NSF GEN in the city center are shown in Figure 7a,b. The averaging was performed for the month in which the counts of the nucleation days (or the f NPF ) was the largest in each year. This was typically in March or April. The curves exhibited a single peak at noon. The baseline of the peaks from 00:00 to 07:00 and from 19:00 to 24:00 was around unity. In some years, the baseline of the NSF NUC was elevated which can be explained by the fact that the particle growth can take place until the late morning of the next day; thus, the NPF can influence the N 6-100 concentrations even in the next morning. Furthermore, we can observe that the peaks have a longer tail in the afternoon side due to the particle growth process. The elevated baseline is a real effect, and it should be included in deriving the mean NSF NUC . The contribution of NPF to the regional concentration level (N 100-1000 ) is the largest at noon, when it reaches a factor of 2.0 to 3.5 (typically 2.5). This is a considerable enhancement, although it only lasts for a few hours. The differences among the annual mean values were likely caused by the year-to-year variability. The diurnal variation in N GEN also exhibited a single peak with a maximum at noon and with a tail in the early afternoon. The maximum values represented concentration contributions from 50% to 100% due to NPF event for a limited time interval. This means that NPF events have an important contribution to UF particles during the midday even in the city center. The values are lower estimates since a considerable part of the N 100-1000 particles can be also produced by NPF from previous days. It is informative to compare the contribution values to the global share of various source sectors in primary UF particle number emission to have an idea on the relative extent of our results. Road transport, power production, and residential heating combustion are the first three largest contributors to primary UF particles with the shares of 40%, 20%, and 17%, respectively [24]. The actual contributions can vary in different parts of the world and with economic development.

Summary and Concluding Remarks
Particulate matter is one of the criteria or key air pollutants (if not the most relevant species) which represents the largest health risk for humans in the world. Its monitoring, quantification, and legislation are based on PM mass. It is, however, increasingly recognized that particle number concentrations in urban and industrial areas are important and valuable additional metrics both for health risk and some environmental considerations. As a consequence, the particle number concentrations have been proposed to amend the air pollutants monitored at present. We showed here that the total number of particles and, in particular, the concentrations in several size fractions, are very useful quantities for these goals. The former property can easily be measured for long-term monitoring purposes, while the latter quantities are more advised to be determined as part of background research studies preparing the actual monitoring activity and in an occasional or expedient manner than for continuous observations. This can create a nice example for mutually beneficial and close cooperation among researchers and regulatory bodies.
Ultrafine particles make up a very considerable portion, typically 70-80% of total particles. The atmospheric residence time of these particles is relatively short, and their concentrations can change rapidly in time and space. Therefore, they reflect the active source processes, atmospheric transformations, and sinks of particles in a dynamic way. This is an important advantage of particle numbers with respect to PM mass as air quality metrics.
Particle number characteristics including the concentration levels (annual medians of approximately 9 × 10 3 cm -3 ), time trends (no strong seasonal dependency), and diurnal variations (with a remarkable temporal pattern) in Budapest are similar to those in most continental large European cities. Their main source types include vehicle road traffic (and other high-temperature emission sources such as household and residential heating, cooking), and atmospheric NPF. The latter process is usually of regional character, and therefore, it seems to be better assessable for contribution or quantification. It was demonstrated that the occurrence frequency of NPF events is considerable even in large cities with an annual mean value of ca. 20%. We also estimated that NPF events increase the ratio of UF particles with respect to the regional particle numbers (N 100-1000 ) by 12% in the city center and 37% in the near-city background. At the same time, the pre-existing UF concentrations are doubled on the NPF event days. The contributions exhibit substantial diurnal and seasonal variations.  Data Availability Statement: The observational data are available from Imre Salma upon reasonable request.