Characterization of Background Aerosol Properties during a Wintertime Smog Episode

The aim of this paper is to study the wintertime physical properties of atmospheric aerosol particles on the basis of data observed at the K-puszta regional background station in Hungary. In Hungary wintertime smog episodes are linked to strong stable air (high pressure blocking events) with thermal inversion. These atmospheric conditions are frequently formed during winter months (November–February) due to the special geographical location of the country. The formation of smog events is highly probable in cases of thermal inversion periods sustaining for at least 4 days. We discuss in the paper the role of high-pressure blocking events in aerosol properties in terms of PM10 concentrations, aerosol size distributions, new particle formation and optical properties. We found that high-pressure blocking events have significant impacts on the size distribution and particle formation processes. At K-puszta the aerosol is in highly aged state with size distribution dominated by the accumulation mode. This is further supported by the optical properties, e.g., by high scattering Ångstrom exponent and by relatively weak absorption. The most significant effect of extreme episodes is manifested in the changes in PM10 concentrations and, consequently, in aerosol optical properties. The PM10 concentrations, scattering coefficients and absorption coefficients considerably increase to extreme values that are characteristic of a heavily polluted atmosphere rather than rural air. Our results indicate that in winter, the air quality at Kpuszta is often influenced by regional air pollution as shown by spatial distribution of PM10 concentration. It is found that PM10 had almost the same concentration in regional background air and in different types of urban environments. The special meteorological conditions and the role of regional-scale transport can explain why local abatements in cities cannot lead to significant improvement of the air quality during smog events.


INTRODUCTION
Important air pollution episodes regularly occur during the winter months.Deterioration of air quality is receiving attention in urban environments because it plays an important role in our everyday life, from public health issues to the economy.Elevated levels of air pollutants directly and indirectly affect human health, aerosol optical properties and visual range.The health impacts of air pollution are mainly associated with respiratory and cardiovascular diseases, which lead to higher mortality and morbidity in particularly vulnerable populations.
It is well established that air quality and weather events are interrelated: meteorological situations frequently lead to negative impacts on urban and rural air quality (see London Great Smog in 1952 or photochemical smog in large cities, such as Los Angeles).From the first appearance of great London smog until recent days, wintertime smog is a well-known phenomenon affecting urban, background and remote environments all over the world.(e.g., Ram et al., 2012;Zotter et al., 2013;Guo et al., 2014;Jiang et al., 2015).
Air quality depends on a number of factors, most notably, the sources of primary air pollutants.In addition, meteorology can also significantly contribute to the formation of smog episodes, which can be further aggravated by unfavorable topography (i.e., mountainous areas and basins).Severe urban air pollution episodes can form in winter and persist for several days to weeks.Smog events are closely related to atmospheric stability and generally form during cold temperatures and weak winds (e.g., Guo et al., 2014;Jiang et al., 2015;Yuan et al., 2015;Zheng et al., 2015).Under strong stable air, horizontal and vertical mixing are limited due to the development of a thermal inversion layer; consequently, the concentration of atmospheric constituents can be extremely high.
In Hungary the formation of strong stable atmospheric conditions is strongly affected by the geographical location of the country.Hungary can be found in a basin surrounded by the Carpathians, the Alps and the Dinaric Alps.Generally due to a cold front the total space of the Carpathian basin is filled with cold air.After the passage of the cold front an anticyclone builds up with lessening winds and calm weather.At night the surface strongly cools down as a consequence of clear sky and intensive surface radiation (especially in case of snow cover); and a quickly growing radiation inversion is formed above the surface that generally extends to the whole Carpathian basin with an average height of 300-600 m (Cséki, 2010).
These atmospheric conditions frequently formed during winter months.According to a comprehensive study  the number of wintertime thermal inversion over Hungary sustaining for 4 or more days is rather high, they occur 6 times per year on the average (Cséki, 2010).It has to be mentioned that the formation of smog episodes are linked to these longer periods.In the last years (2007)(2008)(2009)(2010)(2011)(2012) the wintertime inversion periods still occurred regularly which is shown in Table .1.Each wintertime during the 6 years inversion periods have been formed about 4-7 times, the longest period lasted 33 days.During these periods the PM 10 well describes the scale of air quality.
The wintertime pollution episodes are normally linked to a number of air pollutants like aerosol mass concentration (PM) and trace gases.Over the past few decades, the emission and concentration of SO 2 have decreased significantly in most European countries; however, bad air quality due to the occurrence of winter-type smog episodes is still a serious problem in several European cities (Richter and Peter Williams, 1998).For example, episodes of high ambient PM 10 concentrations (exceeding the EU limit value of 50 µg m -3 ) occur regularly in Ostrava (Czech Republic) (Baranova and Hovorka, 2013), Krakow (Poland) (Larsen et al., 2008) and Budapest, Hungary in winter as indicated by smog alerts of the recent years.This means that due to the emission mitigation policy, among the air pollutants PM plays a major role in these episodes.In Hungary under wintertime smog events the concentrations of SO 2 and NO 2 increase similarly to other observations obtained e.g., in China, Asia (Jiang et al., 2015), however, unlike PM 10 , their concentration only rarely exceeds the 24-hours health standards.Recently, winter smog episodes have been mainly characterized by the high exposure of particulate matter.Fine fraction of PM obviously plays a very important role in haze and smog formation.The contribution of secondary species (e.g., organic compounds, sulfate, nitrate, and ammonium) in fine PM is generally high suggesting the significance of secondary aerosol formation most likely via heterogeneous reactions (Jiang et al., 2015;Zheng et al., 2015).The mass concentration of aerosol particles, mainly in the fine fraction is influenced by long-range transport (Richter and Peter Williams, 1998;Ferenczi, 2013).According to EEA (2013), in several European countries significant fractions of PM 2.5 (sometimes > 50%) can be attributed to long-range transport.As a result, air pollution episodes cannot be confined exclusively to urban air; they also affect the background or rural air quality.Smog areas can even extend to 100-1000 km 2 (Breiling, 1993).
One of the best known manifestations of the long-range transport of anthropogenic air pollutants is Arctic haze.Arctic haze is mainly a result of increased anthropogenic aerosol loads from northern Europe and northern Russia.During winter and early spring, a high concentration of accumulation mode particles can be measured in the Arctic.The episode may last from a few days to weeks when strong surface-based temperature inversion forms, which is characterized by a cold, stable atmosphere that inhibits vertical mixing and the formation of cloud systems and precipitation (Quinn et al., 2007).
The main focus of this work is to study the different aerosol parameters under a typical wintertime smog episode in Carpathian basin.Between 2007 and 2012 there were several sustained smog periods when the air pollution built up; and at the beginning of 2010 a quite long period occurred which was highly appropriate for a general characterization of wintertime smog periods.The changes in aerosol number and PM 10 concentrations, the aerosol size distribution and the aerosol optical properties are discussed as a function of the meteorological conditions.

Site Description
All of the measurements were conducted at the K-puszta station in 2010.The K-puszta background air pollution monitoring station is a rural measurement site located in Hungary (46°58'N, 19°35'E, 125 m a.s.l.) on the Great Hungarian Plain, 15 km northwest of the nearest town of Kecskemét and 70 km southeast of Budapest.The station belongs to the Global Atmosphere Watch (GAW) and European Monitoring and Evaluation Programme (EMEP)

DMPS Measurements
The aerosol number size distribution is monitored by a DMPS (Differential Mobility Particle Sizer) (Aalto et al., 2001); the size distributions of aerosol particles in the 5-800 nm size range are measured approximately every 10 minutes.The particles are classified according to their electrical mobility; then the number concentration of the particles in different size classes is measured by a condensation particle counter (TSI3772).
The aerosols are sampled at a height of two meters and one meter from the station wall.No special inlet is used (only a rain cover).The sampling line is 4 mm diameter copper tubing.Ni-63 370 MBq aerosol neutralizer is applied, and the aerosol classifier is a copy of a Hauke-type 28 cm long DMA.The aerosols are dried with a Topas aerosol drier.Because a drier is used, the size distribution of dry particles is measured.

Aerosol Optics
The optical properties of aerosol particles < 10 µm are discussed.The NOAA-type sampling inlet is 10 m above the surface.The sampled air is dried prior to entering the instruments using a Permapure heatless dryer by diluting the dry particle-free air; the relative humidity of the sampled air is kept under 50%.
The scattering and backscattering coefficients are measured by an integrating nephelometer (TSI3565) at wavelengths of 450, 550 and 700 nm.The averaging time and the sampling rate are 1 min and 29.5 L min -1 , respectively.The nephelometer measures the light intensity scattered between 7° and 170°.The scattering coefficients are corrected for truncation error according to Anderson and Ogren (1998).Regarding the 1 min averaging time, the detection limits of the scattering coefficients at 450, 550 and 700 nm are 0.65, 0.25 and 0.38 Mm -1 , respectively (Anderson et al., 1996).A span gas check (filtered air (low span) and CO 2 (high span)) is conducted monthly, and zero background check is performed every hour.
The aerosol light absorption is measured by a Particle Soot Absorption Photometer (PSAP).This instrument works at 565 nm wavelength, which is extrapolated to 550 nm using the Ångström exponent (α) of -1 (Müller et al., 2011).In the PSAP, a filter-based method is used to measure the absorption coefficient (Bond et al., 1999).The transmittance and the derived absorption coefficient depend on the particle load of the filter, the light scattering of the particles deposited on the filter, the flow rate and the spot size.The absorption coefficients are corrected for these parameters (Bond et al., 1999).When the filter transmittance reaches 70%, the filter is changed; in cases where the transmittance is below 70%, the absorption coefficients are disregarded.In the case of the PSAP, the flow rate and the averaging time is 0.3 L min -1 and 1 min, respectively.This instrument was involved in an intercomparison of aerosol absorption photometers in 2005 and 2007 (Müller et al., 2011).

PM 10 and Meteorology
A standard automatic meteorological station, which is operated by the Hungarian Meteorological Service, is installed at the sampling site.The surface (i.e., 2 m height) temperature, relative humidity, air pressure, global radiation, precipitation and wind speed are measured in 10-minute intervals.Similarly, PM 10 concentrations are also available with a 30-minute time resolution.At both sampling sites PM 10 were measured by a beta-gauge particulate monitor.Finally, the height of the planetary boundary layer is calculated every 3 hours by the ECMWF (The European Centre for Medium-Range Weather Forecasts).

Classifying Particle Formation Events
Particle formation events are classified on the basis of daily (00:00-24:00) DMPS spectra according to the classification scheme recommended by Yli-Juuti et al. (2009).The classification is performed visually by a three-person group to minimize the subjectivity.The days are classified into events, non-events, or undefined data groups.On event days, particle formations are intense and a distinctly new particle mode (i.e., nucleation mode; < 25 nm) is observed (Dal Maso et al., 2005).Non-event days are generally characterized by a bimodal size distribution in which particles < 25 nm are lacking.On undefined days, particles < 25 nm are often present but without a distinct new particle mode.The summary of the event classification for the study period is presented in Table 2.

Size Distribution
The aerosol size distribution is discussed in terms of the modal particle size of nucleation, Aitken and accumulation size ranges.The modal particle diameters are calculated as (Eq.( 1)): , , , where N, Ai and Ac denote the nucleation, Aitken and accumulation modes, respectively.The cut-off diameters of 25 and 100 nm are applied according to the recommendation of Dal Maso et al. (2005).The total mean particle size (TD) is calculated as above but over the entire 5-800 nm size

Derived Intensive Optical Properties: Single Scattering Albedo, Ångström Exponent, and Hemispheric Backscattering Ratio
Additionally, intensive optical properties, such as singlescattering albedo (at 550 nm), the backscattering fractions (at wavelengths of 450, 550 and 700 nm), and the Ångström exponents for the total and backscattering coefficients are discussed.The single scattering albedo (ω 0 ) is defined as the ratio of the scattering and the extinction (scattering + absorption) coefficients.The backscattering coefficient (σ bsc,λ ) measured by the nephelometer is the scattered light intensity in the backward hemisphere of the particle (90°-180°).The hemispheric backscatter (b λ ) fraction is the ratio of the backscattering coefficient and the total scattering (e.g., Seinfeld and Pandis, 1998).
The scattering Ångström exponent is an efficient way to study the aerosol size distribution and obtain information on air pollution as first proposed by Junge (1963) because it is inversely related to particle size.For aerosols dominated by submicrometer particles, the scattering Ångström exponent is large, while for coarse particles, the value is low (e.g., Pereira et al., 2011).The scattering Ångström exponent is calculated as (Eq.( 3)):

Meteorology
By 13th of January of 2010 an anticyclone extended over the Carpathian Basin and the impact of high pressure weather system was dominant in the next weeks.The thermal structure of the atmosphere was studied by using the radiosonde records carried out in Budapest (Budapest, UWYO, 2014).The inversion evolved in one day -over the 900 hPa level -and parallel to this PM 10 concentration increased day by day.The height of the inversion layer changed according to the variation of the weather situation, but the regular attendance of the inversion was found in the lower troposphere till the middle of February.Because of the lack of intensive fronts, a persistent air pollution episode occurred, which greatly affected the air quality over K-puszta.During the smog event the ambient temperature significantly decreased, along with weak easterly surface winds (see Table 3).Owing to cold air pool the mixing layer depth was very low (240 m on average) which favors the development of extreme air pollution.In Fig. 1 the European weather situation is shown at the beginning of the smog period and on January 28, 2010 when the highest PM 10 was measured at K-puszta.One can see on both maps that an anticyclone over the East-European Plain and Central-Europe played decisive role in the control of the weather in Hungary.

PM 10 , Number Concentration and Number Size Distribution
During the entire study period (January-February, 2010), the daily average PM 10 concentration was rather high (Table 4) and reached the 24-hour threshold limit (50 µgm -3 ).During the smog, the daily average PM 10 (60.5 µgm -3 ) was nearly twice as high as the concentration prior to and after the episode (32.0 µg m -3 ) (Table 4).As shown in Fig. 2, in several cases, the aerosol mass concentration briefly exceeded 100 µg m -3 .Such high PM 10 concentrations generally characterize plumes of biomass burning or other smoky air masses (Garland et al., 2008).In some cases, the PM 10 shows low concentration, probably due to favorable weather conditions (e.g., higher wind speed).
The average aerosol number concentration (TC) was 7115 particles cm -3 (in the 5-800 nm size range).Although the average number concentration was higher in the smog (7355 cm -3 ) than in the non-smog period (6501 cm -3 ), this higher number concentration could be attributed to the inhibited vertical mixing and weak aerosol removal processes (cloud and precipitation removal), as seen with Arctic haze episodes (Quinn et al., 2007).Generally, in Hungary, winter is the least active period in terms of new particle formation (NPF) processes (Bécsi et al., 2013) due to unfavorable atmospheric conditions and the low biological activity of the surrounding vegetation.The lowest NPF rate was found during the smog episode (0.054 event/day), while beyond this period the formation rate was three times higher (0.182 event/day).The inactivity of nucleation processes is further supported by the low percentage of nucleation mode aerosols (N% = 5%, see Table 4).The wintertime aerosols at K-puszta were rather aged, as shown by the relationship between the accumulation mode number concentration and PM 10 (Table 5), i.e., PM 10 is generally controlled by the accumulation mode particles.As shown in Figs. 3 (a) and 3(b), the average number size distributions in the smog and non-smog periods were multimodal, with modes in the Aitken and accumulation size ranges.Under smog, the number concentration was dominated by the accumulation mode (Ac% = 53%) with a 41% contribution of the Aitken mode.The deconvolution of the number size distribution showed a relatively small peak in the Aitken mode at 53 nm.Significant modes were found at 103 nm and 165 nm.The former presumably contained particles both from the Aitken and the accumulation size ranges as the results of coagulation and condensation processes.Beside the peak at 165 nm a non-negligible mode was also found at 278 These all resulted in a total mean diameter (TD) of 133 nm.The number size distribution suggests an aged aerosol at K-puszta, contrary to the results obtained in Reno, Nevada, where the elevated PM 10 corresponded to freshly emitted aerosols (Gyawali et al., 2012).Larsen et al. (2008) also concluded that in the region of Krakow (Poland), the combination of coal combustion, traffic exhaust and the atmospheric conditions (low winds, temperature inversion) could be blamed for similar pollution events.Contrary, Baranova and Hovorka (2013) found that in Ostrava, Czech Republic the average number size distribution during smog episodes was nearly monomodal with a mode at 150 nm.
Beyond the smog, the TD decreased by 21 nm (see Table 4) as a consequence of the modified shares of the Aitken and accumulation modes (48% and 44%, respectively).The number size distribution (Fig. 3(b)) was characterized by a mode at 62 nm in the Aitken size range, and three other peaks in the accumulation mode (at 130 nm, 225 nm and 382 nm).In Ostrava, Baranova and Hovorka (2013) also observed a bimodal size distribution after the pollution episode; however their results substantially differed from the size distribution obtained at K-puszta.They found a first mode at 25 nm and a second at 55 nm.The mode at 25 nm indicates * PM 10 total concentration * a Dark and light gray indicate positive and negative (inverse) (p > 0.95), while "*"denote weaker (p > 0.90) relationships, resp.(T: total, N: nucleation mode, Ai: Aitken mode, Ac: accumulation mode; 1 scattering at 550 nm; 2 absorption at 550 nm; 3 back scattering at 550 nm; 4 single-scattering albedo at 550 nm; 5 scattering Ångström exponent; 6 backscattering coefficient ; 7 planetary boundary layer).the importance of freshly emitted particles.Contrary, at Kpuszta the peak centered at 62 nm indicated that fresh particle formation was not significant; and the modes in the accumulation size range also refer to aged aerosol.
The absorption coefficients also increased during the smog episode (see Table 5), however, these values (16.1 and 11.0 Mm -1 in the smog episode and beyond this period, respectively) were significantly lower than the values observed in large cities.In urban air (Reno, Nevada) under winter smog conditions, the scattering and absorption coefficients were nearly twice as high as those on clean days.Estimated from Fig. 6 in Gyawali et al. (2012) in Reno the scattering coefficients (at 355, 405, 532, 870nm) were 80, 45, 35, 20 Mm -1 (clean days) and 135, 90, 60, 30 Mm -1 (polluted days), respectively; while the corresponding values for absorption coefficients were 12.5, 10, 7.5, 5 Mm -1 (clean days) and 30, 25, 17, 12 Mm -1 (polluted days).Our results showed a similar enhancement of the scattering coefficient, while the increase in the absorption coefficients during the smog was much smaller (32%).Gyawali et al. (2012) concluded that the doubling of the absorption coefficient was due to direct traffic emissions.At K-puszta, the lower enhancement of the light absorption may imply that the sampling site is not directly affected by traffic or other emission sources.
Because scattering and absorption coefficients are direct functions of the aerosol (number and mass) concentration, both coefficients had statistically strong linear relationships with PM 10 , total (TC) and accumulation mode (AcC) number concentrations (Table 5).Using linear regression, 3.5 m 2 g -1 and 0.27 m 2 g -1 average mass scattering and absorption efficiencies (at λ = 550 Mm -1 ), respectively, were obtained.This mass scattering efficiency is somewhat higher than the values given by, Garland et al. (2008) (2.8 m 2 g -1 for Guangzhou, Southeast China, referring to summer of 2006) and Virkkula et al. (2011) (2.6 m g -1 in Hyytiälä, Finland); however, it is within the range given by Malm and Hand (2007) at IMPROVE sites.In urban environments higher mass scattering and absorption coefficients could be obtained, e.g., fall of 2006 in urban Guangzhou Andreae et al. (2008) obtained average values of 4.27 and 0.79 m 2 g -1 , respectively.At 450 nm and 700 nm, the average scattering efficiencies were 4.8 and 2.2 m 2 g -1 , respectively, which are also higher than those obtained for Hyytiälä, Finland (3.7 m 2 g -1 and 1.7 m 2 g -1 , respectively) (Virkkula et al., 2011).In addition, the total scattering and absorption coefficients are direct functions of modal diameters (TD, AiD and AcD) and clearly indicate the well-known relationship between optical properties and particle sizes.
The average hemispheric backscatter fraction increased with wavelength (0.10, 0.11 and 0.15 at 450 nm, 550 nm and 700 nm, respectively) and decreased with increasing PM 10 .b λ inversely depends on the total mean and accumulation modal diameter: a smaller mean diameter corresponds to a higher effectiveness of backscattering and the aerosol size dependency of the scattering phase function (e.g., Seinfeld and Pandis, 1998).During the smog episode, b λ decreased by 8% (450 nm), 9% (550 nm) and 10% (700nm) compared to the non-smog period.
The single scattering albedo (SSA) in January-February 2010 (0.91) was similar to the values obtained for other sites e.g., Virkkula et al. (2011).Under the smog episode, the average SSA increased by 2% which is related to the increase in the PM 10 and the greater increase in the scattering relative to the absorption coefficient.This high SSA generally indicates well-aged regional haze aerosol which is characterized by more scattering and less absorbing properties (e.g., Andreae et al., 2008).A haze event in Leipzig, Germany was attributed to regional transport and consisted of a mixture of urban and Arctic haze aerosols (Müller et al., 2004); the single-scattering albedo was similarly as high (0.97 ± 0.06 at 532 nm) as that observed at K-puszta.
The average Ångström exponent (α) of the total scattering coefficients was above 1.9 for the wavelength pair of 450/700 nm (Table 4).The high values also show that at K-puszta, the aerosol is aged and is generally dominated by fine fraction (Wang et al., 2015).This is further supported by the Ångström exponents of well-aged aerosol (1.8-2.8;wavelength range: 355-532 nm) presented by Müller et al. (2004).During the smog episode, the exponent decreased (see also Wu et al., 2015) as a result of the increase in the modal diameter, which is consistent with the inverse relationship between the Ångström exponent and aerosol size.For backscattering, the wintertime average Ångström exponent was 1.01 (450/700 nm).

The Effect of Meteorological Parameters on Aerosol Properties
The statistical relationship of aerosol properties to meteorological parameters is summarized in Table 5.We found that meteorological parameters mostly influence PM 10 and the number concentration in the accumulation mode (AcC); as a consequence, the aerosol optical properties (scattering and absorption coefficients) are affected.As mentioned, the development of a high-pressure system had an important effect on the PM 10 concentration.Along with rising surface pressure and decreasing temperature and wind speed, a high level of PM 10 was measured.Because PM 10 is mainly governed by the number concentration in the accumulation mode, AcC varies similarly to PM 10 , along with the variation of these meteorological parameters.As the high-pressure system developed, the AcC increased by ~50%.
Variation in the surface pressure, temperature and wind speed was also reflected in the total mean diameter (TD).In addition, higher RH results in larger TD and AiD.The daily variation in the RH and AiD indicate that Aitken mode particles could be hygroscopic.However, the accumulation modal diameter (AcD) is not affected by relative humidity.Generally, particle growth is linked to vapor -including water vapor -that condenses on the particles.Aitken mode particles are significantly smaller than particles in the accumulation mode; because of water vapor condensation, the change in the AiD size is much more distinct than in the accumulation mode.Finally, the nucleation mode diameter (ND) was not noticeably affected by the meteorological parameters, including the global radiation.
Among the modal fractions, N% was positively correlated with global radiation, wind speed and temperature but negatively correlated with RH.Ai% and Ac% varied oppositely as functions of air pressure, temperature and wind, as shown in Table 5.All these variations reflect the development of the high-pressure system.
Because of the changes in the meteorological parameters, the scattering and absorption coefficients varied similarly to the PM 10 and number concentrations.In addition, as the height of the planetary boundary layer lowered, the scattering coefficient increased.The high-pressure buildup decreased the backscatter fraction and the Ångström exponent, most likely due to the growth of the average particle size.Finally, under the high-pressure system, the SSA increased as a result of the greater increase in the scattering coefficient than in the absorption coefficient.
The effect of regional air pollution on the background air quality at K-puszta can be collated by comparing parallel PM 10 concentrations and aerosol extinction coefficients measured in Budapest and K-puszta (Fig. 2).The extinction coefficients in Budapest were derived from visual range observations (e.g., Molnár et al., 2008).The average PM 10 concentration at K-puszta was generally the same as that in Budapest (49.9 and 44.9 µgm -3 , respectively).During the smog, the average PM 10 in Budapest increased by 40% compared to non-smog period, while at K-puszta the average concentration nearly doubled.This parallel variation of wintertime PM 10 concentration in Budapest and at K-puszta is rather common as shown in Table 1 for the periods from December to February between 2007 and 2012.The direct relationship between the enhanced PM 10 concentration and permanent inversion periods resulting occasionally in heavy smog periods is affirmed.
At K-puszta, the average dry extinction (scattering + absorption) coefficient at 550 nm was 189 Mm -1 , while a similar value (255 Mm -1 ) was obtained in Budapest.Under smog a 1.5 times increase in the dry extinction coefficient was observed in Budapest while at K-puszta the corresponding value was 1.9.These data directly imply that air pollution in Hungary (Central Europe) is not confined to the large cities; it actually causes serious problems outside of towns and in rural areas.This is further supported by the spatial distribution of PM 10 during January-February, 2010.In Fig. 4 PM 10 concentrations during the smog and non-smog periods are shown for different parts of Hungary representing mainly urban environment (Hungarian Air Quality Network) and at K-puszta).It has to note that at all sampling sites the PM 10 concentration was significantly (average: 180%, range: 150%-225%) higher during the smog than non-smog period.Despite the greatly different environments during the smog the average PM 10 concentrations were rather comparable at all sites.This similarity directly indicates the importance of regional transport in PM 10 concentration.

Fig. 1 .
Fig. 1.European weather situation (a) at the beginning of the smog period (January 14, 2010.00 UTC) and (b) on January 28, 2010 when the highest PM 10 was measured at K-puszta (H: high pressure, L: low pressure).

Fig. 3 .
Fig. 3. Measured and fitted aerosol number size distribution during winter smog (a) and non-smog (b) periods.(Under smog and non-smog periods the goodness of the fits were r 2 = 0.997 and r 2 = 0.993, respectively).

Fig. 4 .
Fig. 4. Spatial variation of average (2 months) PM 10 concentration in Hungary, January-February, 2010.Data are from Hungarian Air Quality Network and Hungarian Meteorological Service (K-puszta).

Table 1 .
Smog period statistic between 2007 and 2012 at K-puszta and Budapest.

Table 3 .
Summary of averaged meteorological parameters.

Table 4 .
Summary of aerosol characteristics during the entire studied period (January-February, 2010), the smog episode and beyond.

Table 5 .
Statistical relationship of PM 10 , total and modal number concentrations, diameters and modal fractions with aerosol optical characteristics and meteorological parameters a .