a study of aerosol particle sizes in the atmosphere of athens, greece, retrieved from solar spectral...

This article was published in an Elsevier journal. The attached copyis furnished to the author for non-commercial research and

education use, including for instruction at the author’s institution,sharing with colleagues and providing to institution administration.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

A study of aerosol particle sizes in the atmosphere of Athens, Greece,retrieved from solar spectral measurements

A.D. Adamopoulos a,1, H.D. Kambezidis b,⁎, D.G. Kaskaoutis b,2, G. Giavis b,2

a Directorate of Air and Noise Pollution Control, Ministry for the Environment Physical Planning and Public Works,Patission 147, 11251 Athens, Greece

b Atmospheric Research Team, Institute of Environmental Research and Sustainable Development, National Observatory of Athens,P.O. Box 20048, 11810 Athens, Greece

Received 12 September 2006; received in revised form 12 April 2007; accepted 12 April 2007

Abstract

This work aims at determining the aerosol particle radii in the atmosphere of Athens. Such a work is carried out in Athens forthe first time. For this purpose, solar spectral direct-beam irradiance measurements were used in the spectral range 310–575 nm. Toestimate the particle radius from aerosol optical depth retrieval, a minimization technique was employed based on the golden-section search of the difference between experimental and theoretical values of the aerosol optical depth. The necessary Miecomputations were performed based on the algorithm LVEC.

In this study, the mean particle radius of a given distribution was calculated every 30 min during cloudless days in the periodNovember 1996 to September 1997. The largest particles were observed in the summer and the smallest during winter. The resultwas verified by the increased values of the aerosol optical depth and the turbidity factors calculated in the summer. The differencesin the diurnal variation from season to season are attributed to the prevailing wind regime, pollutant emission and sink rates in theatmosphere of Athens.© 2007 Elsevier B.V. All rights reserved.

Keywords: Aerosol optical depth; aerosol size distribution; solar spectral irradiance; Athens; Greece

1. Introduction

Aerosol particles play a key role in air quality andatmospheric physics and chemistry. They interfere withweather and modify climate, participate in cloud devel-opment as condensation nuclei, change the electricalproperties of the atmosphere and interact with atmo-spheric radiation (Charlson et al., 1991). Therefore, theyintervene in the global radiation budget and atmosphericchemical cycles and they also modify the optical prop-erties of the atmosphere (Haywood and Boucher, 2000).The climatic effect of aerosols is closely related to theirphysical properties and size, the surface albedo and the

Atmospheric Research 86 (2007) 194–206www.elsevier.com/locate/atmos

⁎ Corresponding author. Tel.: +30 10 3490119; fax: +30 10 3490113.E-mail addresses: [email protected]

(A.D. Adamopoulos), [email protected] (H.D. Kambezidis),[email protected] (D.G. Kaskaoutis), [email protected](G. Giavis).

URL's: www.meteo.noa.gr/ENG/iersd_art.htm (H.D. Kambezidis),www.meteo.noa.gr/ENG/iersd_art.htm (D.G. Kaskaoutis),www.meteo.noa.gr/ENG/iersd_art.htm (G. Giavis).1 Tel.: +30 10 8650076; fax: +30 10 8645589.2 Tel.: +30 10 3490119; fax: +30 10 3490113.

0169-8095/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.atmosres.2007.04.003


relative altitude between aerosol layer and clouds (Abelet al., 2005; Kinne and Pueschel, 2001). Troposphericaerosol properties are generally associated with micro-physical phenomena and atmospheric dynamics thatdetermine the size distribution and chemical composi-tion of the aerosols.

Aerosol characterization is an important task inaerosol studies using atmospheric optical depth data(D'Almeida et al., 1991; Cachorro and Tanré, 1997;Bates et al., 1998). Many studies deal with particle-sizedistribution, complex index of refraction and other sizeparameters in atmospheric columns worldwide (Dubo-vik et al., 2002; Vitale et al., 2000; Cachorro et al.,2000). With the establishment of the Aerosol RoboticNetwork (AERONET) an effort for global aerosolmonitoring is attempted and their optical and physicalproperties (e.g. size distribution, single scatteringalbedo, asymmetry factor and refractive index) areachieved through the Dubovik and King (2000)algorithm. The size and the shape of the suspendedparticles, in conjunction with their chemical composi-tion, deal with the important scattering and absorptionprocesses in the atmosphere (Molnàr and Mészáros,2001). They found that submicron particles withdiameter b1 μm are responsible for 82% and 7% ofthe scattering and absorption of the solar radiation,respectively.

Due to the variety of the regions surrounding theMediterranean, different types of particles can be foundthroughout the region, having both strong temporal andspatial distribution (Antoine and Nobileau, 2006; Bar-naba and Gobbi, 2004); desert dust, originating fromSahara, polluted aerosols produced mainly in industrial-ized areas of Continental and Eastern Europe, aerosolsformed over the Mediterranean itself or transported fromthe North Atlantic and biomass smoke often producedfrom seasonal forest fires. The various aerosol typesdisplay different optical and physicochemical character-istics. Their properties, such as optical depth, chemicalcomposition, size distribution, single scattering albedo,asymmetry factor and vertical distribution vary largely,thus affecting the radiative and energy balance in differ-ent ways (Horvath et al., 2002; Markowicz et al., 2002).

Furthermore, Greece is one of the most interestingregions for aerosol studies, because soil dust from theneighboring North African desert areas, and the emissionof local aerosols, such as sulfuric, nitric and carbona-ceous particles, is increasing due to rapid urbanization.Therefore, the location of Athens offers a unique oppor-tunity to monitor aerosols from different sources. On theother hand, the identification of a clear aerosol type overAthens is rather impossible due to the variety of sources

and mixing mechanisms. Recent studies focusing on thischaracterization via aerosol interactions with solar radia-tion led to rather contradictory results (Kaskaoutis andKambezidis, 2006; Kaskaoutis et al., 2006b). This workaims at determining the aerosol particle radii in theatmosphere of Athens for the first time. For this purposesolar irradiance measurements covering the periodNovember 1996 to September 1997 were used in con-junction with the LVEC model. The particle radius isinvestigated according to the season, wind regimes andair mass origin.

2. Topography, data collection and methodology

2.1. Topography and climatology

Athens is a Mediterranean city of about 3.5 millionpeople (census of 2001) located in an oblong basin withan area of 450 km2. The main axis of the basin lies in the

Fig. 1. Map of the Athens basin. The vertical lines indicate theresidential areas, while the horizontal ones show the industrial areas.The summits of the mountains are presented together with their altitudein m a.s.l. The dashed line is the 100-m contour, while the dashed-dotted line refers to the 500-m elevation. The locations of themeasuring site (NTUA) and the meteorological station site (NOA) areshown. The Hellinikon Airport that was in operation at the time of themeasurements ceased its operation in March 2001.

195A.D. Adamopoulos et al. / Atmospheric Research 86 (2007) 194–206


NE-SW direction (Fig. 1). High mountains on threesides and the Saronikos Gulf on the south surroundAthens. The mountains that act as physical barriers, thewarm and dry climate in the warm period of the year andthe concentrated pollution sources are responsible forthe deterioration of air quality. Thus, under adversemeteorological conditions (mainly sea breeze and calmconditions), the concentration of air pollutants maysometimes exceed the air quality standards of theEuropean Union and the World Health Organization(DEARTH, 1998). The Athens basin has two openingsat the SW and NE edges, so that the predominant windsand consequently the aeration of the basin are accom-plished in the NE-SW direction.

The synoptic meteorological pattern during thewinter is affected by (a) the Siberian anticyclonecausing polar continental air flow from North-EasternEurope, (b) an anticyclone over Middle and EasternEurope and (c) the low-pressure systems in the WesternMediterranean and Atlantic Ocean, moving from Westto East. These regimes, in conjunction with the higherprecipitation in this season, result in lower aerosol loadand more transparent atmospheric conditions. On theother hand, there are two opposing wind regimes in thewarm period of the year. With very weak synoptic flow,a sea breeze cell is developed favoring the rise of airpollution levels (e.g. Klemm et al., 1998). On the otherhand, the northerly strong “Etesian” wind that blowsevery summer in the eastern part of the Mediterraneanplays a depollution role in the Athens basin (Clappieret al., 2000).

2.2. Data collection

For the purpose of this study, spectral measurementsof solar direct-beam irradiance were performed by ahigh-resolution spectrometer during 42 cloudless daysin the period November 1996 to September 1997. Thespectral measurements were taken with a sampling rateof 30 min on each of the 42 experimental days. Thedates and the number of spectra collected are tabulatedin Table 1. The measurements were continued through-out the day from morning to afternoon, covering the airmass range 1.03–2.31. The formation of clouds on somedays of the experimental period did not allow spectro-radiometric measurements and therefore on some days,especially in the winter, the number of recorded spectrais rather limited (4–8). Note also that in this period thedaylight hours are much fewer (about 7) compared to thesummer ones (about 14). The majority of the spectralmeasurements took place in the summer (251) due to themore stable and sunny atmospheric conditions. In

contrast, a limited data set (5 days with 49 spectra)refers to the autumn, due to the limited period ofmeasurements, since October is not included. Never-theless, the amount of data in each season allows us tocompute the diurnal pattern of both aerosol optical depthand particle size.

The solar spectra lie in the range 310–575 nm andhave a resolution of 0.55 nm. The measuring site was onthe roof of the Department of Electrical Engineering,National Technical University of Athens in the citycenter (φ=37.97°Í, λ=23.72°Å). The experimentalcampaign was carried out using a Passive Pyrheliomet-ric Scanner (hereafter PPS). The term “pyrheliometer” isused here alongside the conventional term indicatingthat the specific instrument always records the direct-beam component of solar radiation. The term “passive”implies that no energy is required by the PPS to recordsolar radiation, while the term “scanner” refers to thePPS spectral scans and simultaneous recordings at allwavelengths in the spectral region considered. Theapparatus consists mainly of a spectrometer (model SD-1000 manufactured by OCEAN OPTICS Inc.) and asun-tracking system especially built for this purpose.The sun tracker of the PPS consists of a tripod with twonormal axes of rotation in azimuth and elevation. Theelectronic circuits of the tracker involve a four-photodiode system and stepping motors with 0.3°accuracy. The photodiode system driven by an elec-tronic controller is fastened onto a small flat platetogether with the spectrometer lenses. To check the sun-tracker efficiency, an alignment system, similar to thaton conventional pyrheliometers, is fixed on the flatsurface. A perfect alignment occurs when the sun spotfalls on a preset white dot. In all measurements, thealignment was ensured. Solar radiation enters the systemthrough two concentric lenses and is guided by twooptical fibers onto a diffraction grating unit situated in a

Table 1Dates of measurements and total number of recorded spectra

Month Date Number of recordedspectra on respective date

November 1996 4, 10, 14, 16 8, 5, 8, 6January 1997 16, 17, 18, 20, 21, 24,

25, 2818, 7, 18, 18, 14, 14,12, 16

February 1997 1, 3, 4, 11 18, 13, 10, 18March 1997 13, 14, 28 18, 13, 12April 1997 15 21May 1997 5, 12, 15, 20, 29 21, 20, 22, 19, 9June 1997 10, 14, 18, 19, 20, 23 11, 4, 22, 22, 9, 17July 1997 1, 2, 11, 16, 18, 19 22, 22, 22, 8, 22, 22August 1997 1, 4, 5, 6 12, 6, 15, 15September 1997 5 22

196 A.D. Adamopoulos et al. / Atmospheric Research 86 (2007) 194–206


constant-temperature environment and then onto a CCDsurface; the individual spectral values (in mV) are thenconverted from analog to digital form by an A/Dconverter and further into solar irradiance units (Wm−2

μm−1) using a PC with appropriate software.To achieve as high accuracy as possible in the

measurements, the PPS was calibrated against a halogenspectral prototype lamp (model S-771 by OPTRONICLABS) at the start and end of the experimental period.The calibration of the instrument in the intermediatetime interval was achieved by a linear interpolation ofthe recorded values between the two calibrationsassuming a linear degradation of the PPS sensitivitywith time. The OCEAN OPTICS software provided iscapable of estimating the electronic noise spectral dis-tribution and extracting it from the signal; to achievethis, one measurement must be carried out in a darkenvironment before sampling so that solar irradiancemeasurements are corrected for the noise effect. Anumber of applications using PPS have appeared in theliterature involving the estimation of the spectral partialand total atmospheric attenuation properties, the total O3

and NO2 columns, the atmospheric turbidity and aerosolstudies (Adamopoulos et al., 2000, 2005; Kambezidiset al., 1997a,b, 2000a,b, 2001a,b, 2005).

2.3. Methodology

The estimate of the particle-size distribution ofaerosols is based on the spectral features of the aerosoloptical depth (here-in-after AOD). The method used forretrieving AOD from the solar direct-beam irradiancemeasurements using the PPS is discussed by Kambezi-dis et al. (2000b, 2005). More recent studies conductedusing data recorded by the PPS (Kaskaoutis et al.,2006a,b) investigate the errors and uncertainties re-vealed from both AOD and Ångström parameters (α andβ) retrieval. Thus, as clearly established in Kaskaoutiset al. (2006a) the above parameters are more accurateunder turbid conditions. The overall uncertainty inspectral AOD calculations was found to vary between0.01 and 0.03 and be comparable to the values reportedby Smirnov et al. (2002). These errors exhibit a clearwavelength dependence with higher uncertainties in theUV band, where the instrumental errors are larger, andunder high zenith angles due to small signal-to-noiseratio. Furthermore, errors and uncertainties in the AODretrieval by subtracting the contribution of Rayleighscattering, ozone and nitrogen dioxide absorption areunderlined in Kaskaoutis et al. (2006b). The estimationof the Ångström parameters was made using the least-squares method covering the spectral range 320–575 nm

(Kambezidis et al., 2001b). The spectral region below320 nm was omitted since both solar irradiance and thesignal-to-noise ratio are considerably low in this spectralband.

The Mie scattering theory relates the AOD, τα, to theparticle-size distribution, η(r), as:

saðkÞ ¼Z r2

r1

pr2Qðn; xÞgðrÞdr ð1Þ

where Q(n,x) is the Mie extinction-efficiency factorassuming spherical particles with radius r and dependson the refractive index, n=nR− jnI (complex numberconsisting of real nR and imaginary nI parts) and theparticle Mie size parameter, x=2πr /λ, with λ being thewavelength in micrometers. The lower and upper limitsof the particle radii are r1 and r2, respectively.

To retrieve the particle-size distribution, η(r), one hasto apply a classical inversion method (Yamamoto andTanaka, 1969; King et al., 1978; Schmid et al., 1997). Inthis study, a minimization technique is used, which isbased on the golden-section search (Cheney and Kincaid,1980) of the difference between experimental and com-puted Mie values. This “pseudo-inversion” algorithm(Cachorro and Salcedo, 1991), called LVEC, has beenused in several relevant studies (Cachorro and Tanré,1997; Cachorro and de Frutos, 1997; Cachorro, 1998;Vergaz et al., 1999; Cachorro et al., 1999, 2000). Thisalgorithm requires a given function for the particle-sizedistribution. The well-known single log-normal functionhas been selected in this work. Thus, two parameters needto be retrieved: the geometrical or mode radius, rg, and thespread of the distribution, σ. Furthermore, to apply the“pseudo-inversion” method there is also a need to input agiven value to Eq. (1) for the refractive index, n. Anormalization procedure applied to the extinction spec-trum permits an account of only the spectral behaviorand hence a determination of rg and σ by the golden-section-search technique. Values of rg are determined byassuming a value of σ=2.5. The choice of this value isdiscussed elsewhere (Cachorro et al., 2000). For therefractive index, there has been considered a complexvalue n=1.5− j0.01, with the real and imaginary partsrepresenting the scattering and absorption effects, respec-tively. Its value is taken from Vergaz et al. (1999)calculated for a Spanish location with same latitude asAthens and almost similar climatic conditions. Charac-teristic values of the refractive index for the main aerosoltypes (urban/industrial, biomass burning, maritime anddesert dust) are reported by Dubovik et al. (2002). Then,the LVEC algorithm is invoked to derive the geometricalor mode radius.



3. Results and discussion

3.1. Seasonal variation of particles size

Before analyzing the variation of particle size ac-cording to the season, Fig. 2 is given regarding the meandiurnal variation of AOD at 500 nm (AOD500) in orderto establish the seasonal differences in the aerosol load.The crosses, squares, circles and triangles denote valuesfor autumn, winter, spring and summer, respectively.Since the diurnal variation of the atmospheric turbidityfor the same period was well established in a previousstudy (Kambezidis et al., 2001b) via the Ångström andSchüepp turbidity parameters, a short discussion isgiven here. As clearly shown in Fig. 2, the AOD500 ishigher in the summer, where the more stable atmo-spheric conditions favor the accumulation of aerosols,and on specific days when there was significant in-fluence of Saharan dust transport. In contrast, the mosttransparent atmospheric conditions take place in wintersince the stronger winds contribute to the cleansing ofthe atmosphere. Regarding the diurnal variation of theAOD500, this is strongly related to the anthropogenicactivities of the city, and as a consequence, it exhibitshigher values around 9:00 LST, the rush hour. In the restof the morning the values of AOD500 continuouslydecrease. A second increase after midday is depicted inthe summer period. On the other hand, the AOD500

diurnal variation is rather negligible in winter andautumn periods. However, the limited number of spectradoes not allow the establishment of safe conclusions.

In Table 2 the mean seasonal values of the AOD500,the α-Ångström in the band 320–575 nm (α320–575) andthe particle radius r are given. Also, values for twoopposing wind regimes (Etesian winds and sea breeze)are given, which are studied in Section 3.2. The particlenumber, which is also retrieved from this method, varieswidely in the measuring period (from 50±30 to 1400±600), also exhibiting a significant day-to-day or hour-to-hour variation. In the literature there is a lack ofpublished values for this parameter, thus any compar-ison of our results cannot be conducted. The meanα320–575 and AOD500 values in spring are comparableto those obtained by Kaskaoutis et al. (2006a,b), whilethe respective autumn values are similar to thosepresented by Giavis et al. (2005) for another timeperiod in Athens.

The AOD data, calculated from all experimentaldays, were used as inputs in the LVEC algorithm. Thestandard deviation between the theoretical and theexperimental AODs, calculated throughout the spectralrange (320–575 nm), was most often found to be lessthan 0.02 and never exceeded the value of 0.04. Thestandard deviation of the radii was found to vary in therange 10−4 to 10−3 μm, as an order of magnitude, whilethe lower and upper limits of the particle-size distribu-tions are 0.001–0.1 μm. Fig. 3 presents the mean diurnalvariation of the radius of the suspended particles overAthens covering the period November 1996 to Septem-ber 1997. The rather limited number of data, especiallyin the autumn, allows for a first approximation of theaerosol radius in the Athens atmosphere. However, itconstitutes the first study focusing on it. It should benoted that the investigation of the aerosol particle size ina vertical column using sun photometers or radiometersis a rather difficult task due to the variety of aerosollayers and types (e.g. urban/industrial aerosols withinthe boundary layer and a dust plume at an upperatmospheric level). It has been shown that this is themost likely scenario over Athens, especially in the warmperiod of the year (Papayannis et al., 2005).

Fig. 2. Mean diurnal variation of AOD at 500 nm in the four seasons(□ winter, ○ spring, Δ summer and + autumn) over Athens.

Table 2Mean values of aerosol optical depth at 500 nm (AOD500), α320–575,and particle radius, r (in nm), for each season and wind regime

AOD500 α320–575 r

Winter 0.19±0.05 2.15±0.28 8.15±3.80Spring 0.35±0.12 1.04±0.42 27.27±20.46Summer 0.51±0.16 0.69±0.25 37.41±12.12Autumn 0.23±0.02 1.25±0.20 22.37±7.96Etesian winds 0.42±0.21 0.87±0.37 27.90±14.36Sea breeze 0.52±0.14 0.79±0.33 33.72±14.29

The standard deviations of the mean values are also given.



The particle radius achieves its highest levels insummer (37.41±12.12) and the lowest in winter (8.15±3.80) in accordance with the lowest levels in summer(0.69±0.25) and highest in winter (2.15±0.28) of theÅngström wavelength exponent α. The highest valuesof radius of the year, however, are observed duringspring early in the morning, in line with the minimum αvalues (Kambezidis et al., 2001b). The calculated springvalues, excluding the morning ones, are comparable tothose in the autumn and lie between the summer andwinter values. The lowest values of radius, ∼8 nm,which are observed in the winter are in line with thepeaks of α320–575, indicating that the aerosols consists ofvery fine particles; therefore, the dominant solar irra-diance attenuation in the atmosphere of Athens ismainly accomplished by the accumulation of particleswith various sizes. This happens because the giant par-ticles are removed by the dry and wet depositionmechanisms in the winter. In contrast, during summer,the increasing dust dispersion from the neighboringlandscapes or the long-range dust transport from Sahara,in conjunction with the atmospheric stability and theabsence of rainfall, increase the residence time of thecoarse-mode particles in the Athens area. It is helpful tobe reminded here that the small particles are of anthro-pogenic or industrial origin and mainly constitute com-bustion products, i.e. soot, sulfates and nitrates (Yaaqub,1989). These small particles of anthropogenic originattenuate solar irradiance mostly in the ultraviolet andvisible region of the spectrum; they therefore signifi-cantly enhance AOD at these wavelengths. On the otherhand, the coarse-mode particles are of natural origin (seaspray or desert dust) and they attenuate almost equally

the solar radiation both in the short and long wave-lengths (Reid et al., 1999; Schuster et al., 2006). In thepresent study, the particle size was assessed by means ofattenuation of the solar direct-beam irradiance in aspecific spectral region resulting in a limited range ofradius values that varied in the range 0.001–0.1 μmaccording to Mie theory. Another study based on chemi-cal analysis showed that the particles with radius in therange 0.001–0.1 μm (Aitken diameters) in the Thrias-sion Plain (west of the Athens basin) are most likely tobe attributed to lead, nickel, zinc and less to chromium,manganese, sulphates and nitrates (Christides, 1995,p. 108). The Thriassion Plain is a highly industrializedarea and is separated from the Athens basin by Mt.Aegaleo (468 m a.m.s.l.). Nevertheless, under the favor-able meteorological conditions of sea breeze or lightwesterly winds, Asimakopoulos et al. (1992) haveshown that polluted air masses from the Thriassion Plaincan influence the air quality of the Athens basin.

As far as the diurnal variation of the particle radius isconcerned (Fig. 3), no significant variation is observedin the autumn. However, the limited number of spectraldata (47) does not allow for a derivation of safe con-clusions. During the winter, the particle radius appearsto be almost constant throughout the day (10:00–16:00 h LST). The early morning and late afternoonvalues were rejected since the signal-to-noise ratio of thePPS was significantly low due to the low sun elevation.Also, for high solar zenith angles or low radiation, theuncertainties in particle radius retrievals increasedsignificantly. As found in Kambezidis et al. (2001b),the α-Ångström was almost constant in spring takingvalues close to 1.3 that represents a natural and well-mixed atmosphere. This is consistent with the findingsof this study where the particle radius lies in the narrowrange 0.012–0.020 μm. Before 10:30 h LST, highervalues of radius were observed in accordance with thelow α320–575 values. This is attributed to the low mixing-layer height due to the morning surface-temperatureinversion, frequently observed inMay, when the majorityof spring spectra were taken.

A significant role in the increase of particle radius inthe early morning hours is probably the humidificationprocess and the absorption of water from water-solubleurban aerosols (Day and Malm, 2001). Measurements offine-mode (pollution) aerosols at the coasts of Korea andJapan during ACE-Asia with aircraft and ships showedthat the pollution aerosols are moderately-to-stronglyhygroscopic (Carrico et al., 2003). The relative humidity(RH) is a determining factor for the size of the aerosolparticles, especially for the water-soluble industrialparticles, as shown in Seinfeld and Pandis (1997) and

Fig. 3. Mean diurnal variation of the particle radius in the four seasons(□ winter, ○ spring, Δ summer and + autumn) over Athens.



from extensive measurement campaigns (e.g. TAR-FOX). Therefore, some of the particle growth may be theresult of absorption by water vapor, especially sinceAOD and water vapor content (WVC) are highlycorrelated. However, this is very uncertain, since thereis no direct relationship between WVC and RH, while itis not knownwhether RH reaches sufficiently high levelsin order for the aerosol growth to be due to humidifi-cation. The hygroscopic growth at high RH values alsoincreases AOD, due to the enlargement of the accumu-lation-mode particles, such as sulfates, which areproduced in coal combustion (Ferrare et al., 2000). Inthe literature there are some studies (e.g. Hänel, 1976;Horvath, 1996; Smirnov et al., 2002), which correlateRH with the aerosol load and particle size for differentaerosol types. In addition, the low mixing height in theearly morning hours favors the accumulation of aerosolsand, as a consequence, the processes of coagulation,condensation and gas-to-particle conversion becomemore effective.

The temperature inversion disappears after sunrise.More specifically, the highest values of particle radiuswere observed before 9:00 h LST. This coincides withthe rush hours in the city center (Kambezidis et al.,1995; DEARTH, 1998). Except for the morning peak inspring, the highest values of the day for the wholeexperimental period appear in summer. The summerpeak in particle radius shows up at 8:30–9:00 h LST dueto the reasons discussed above. The atmospheric particleradius in a column over Athens is assessed for the firsttime in Athens. This implies that any comparison of theresults to others is not feasible due to the absence ofrelevant studies. From the whole analysis it is concludedthat the mean and diurnal patterns of both AOD andparticle size in the four seasons depend strongly on themeteorological field, the emission rates, the removalprocesses and the mixing mechanisms. In the next sec-tions, a more detailed analysis taking into account thelocal wind-circulation patterns and the long-range trans-port is conducted.

3.2. The key role of sea breeze on particles size

A specific analysis with the spectral data gathered inthe period May to September 1997 (364 spectral datasets) was carried out. The data were sorted into twogroups. The first group took into account data calculatedon days under the dominance of the N-NE Etesian windswith speeds greater than 5 m s−1, while the secondgroup included days of pure sea breeze development (S-SW directions). All necessary wind data for this studywere taken from the inventory of the National Obser-

vatory of Athens (NOA), which is quite close to the siteof the spectral measurements.

The calculated particle radius (Fig. 3) was found tobe smaller on days with Etesian flow (symbol: ×)compared to that under fully developed sea breeze(symbol: ⋄). This indicates that small particles areabundant in the atmosphere of Athens under the pres-ence of strong Etesian winds, while the large particlesare advected southwards to the sea or deposited onto theground. The roles of these two competing wind regimesare confirmed by other studies too (Suppan et al., 1998;Klemm et al., 1998).

During the pure sea breeze days, the synoptic windflow was weak favoring the development of a sea breezecell from the Saronikos Gulf. The calculated transmit-tances due to aerosol scattering were very low becauseof the appearance of the usual air pollution episodes,thus enhancing both Ångström and Schüepp turbidityparameters (Kambezidis et al., 2001b). The results arealso supported by high levels of ground-based air-pollutant measurements performed on a 24-h basis bythe Hellenic Ministry for the Environment (DEARTH,1998) and are consistent with the reduction in visibilityobserved above the city center. This is mainly due toautomobile-originated aerosols. Their influence uponatmospheric turbidity (Kambezidis et al., 2001b), totalatmospheric transmittance (Kambezidis et al., 2000a;Adamopoulos et al., 2005) and aerosol transmittance(Kambezidis et al., 2000b, 2005) has also been inves-tigated over Athens for the same period. The relevantstudies have shown that the atmospheric turbidity ishigher during the summer months, even more on dayscharacterized by sea breeze and especially during themorning hours. The morning peak of the aerosol loadcoincides well with the rush hour of the city and has alsobeen observed in different periods (Kambezidis et al.,1995). The weak synoptic wind flow favors the ap-pearance of a temperature inversion with a low mixing-layer height. As a result, a substantial aerosol burden istrapped at low altitude, the origin of which must be thesoil and the sea spray (Scheff and Valiozis, 1990;Eleftheriadis et al., 1998). Later in the day, the particleradius turns out to exhibit a decreasing trend (Fig. 4).Two secondary peaks are observed; one centered at11:00 h LST and another at 15:00 h LST that coincidewith the time the air masses driven by the sea breezecirculation are advected over the city center. Duringsuch days, the smaller particles are well swept by the seabreeze northwards, while the larger ones are depositedonto the ground. This is supported by the results of otherstudies that show high levels of secondary photochem-ical-pollutant concentration at the foothills of Mts.



Parnitha and Penteli in the north of the Athens basin inthe early afternoon hours (Kambezidis et al., 1998a).There, a convergence zone develops and air masses arelifted to an altitude of almost 2 km from where they cantake part in long-range transport by the synoptic windflow (Klemm et al., 1998). Furthermore, the accumu-lation of aerosols under the favorable conditions of seabreeze circulation may be crucial for their growth due tothe condensation, coagulation and gas-to-particle con-version processes. In addition, the S-SW wind sector isthe preferable direction for the arrival of air masses fromthe Sahara desert (Papayannis et al., 2005).

In contrast, the effective particle radii during theEtesian days have smaller values (Fig. 4). This wasanticipated since the Etesian winds are responsible forcleansing the atmosphere in the Athens basin (Suppanet al., 1998). The results are in accordance with the highvisibility observed and the low levels of air-pollutantconcentration recorded by the relevant network operatedby DEARTH (1998). Similar results are also supportedby other studies that have derived Linke, Unsworth-Monteith, Ångström and Schüepp turbidity parametersand have shown much higher turbidity values during seabreeze days than under Etesian flows (Kambezidis et al.,1998b, 2001b). A peak in particle radius is observed inthe morning, as was the case on the sea breeze days. Onthe other hand, the relative humidity during Etesian daysis lower compared to that during sea breeze days dueto the stronger and drier winds. Probably, this explainsthe smaller particle sizes in the early morning hoursof the Etesian days. After the morning maximum a rapid

decrease of particle radius is observed. As the dayprogresses the particle radius generally undergoes adecreasing trend.

The particle radius computed through the LVECalgorithm was found to be well correlated with the α-Ångström determined in the spectral band 320–575 nm(Kambezidis et al., 2001b). As clearly established byKaskaoutis and Kambezidis (2006), the Ångströmwavelength exponent exhibits strong wavelength de-pendence. Also, Schuster et al. (2006) suggested that thedetermination of α at shorter wavelengths providesinformation about the particle size, while its determina-tion at longer wavelengths is more sensitive to changesin the fine-coarse fraction. Furthermore, Reid et al.(1999) found that the particle diameter is highly corre-lated with α, when its determination takes place atshorter wavelengths. In cases of spectral range expandedto longer wavelengths (NIR band) this correlationbecomes nearly neutral. In the present study the correla-tions between α320–575 (denoted as α for simplicity) andparticle radius r, give the following linear trends:

r ¼ �3:17F1:4aþ 13:7F3:1; R2

¼ 0:33 ðwinterÞ ð2Þ

r ¼ �46:35F2:9aþ 75:5F3:3; R2

¼ 0:92 ðspringÞ ð3Þ

r ¼ �43:8F3:7aþ 67:8F3:1; R2

¼ 0:87 ðsummerÞ ð4Þ

r ¼ �31:9F9:2aþ 62:1F11:6; R2

¼ 0:67 ðautumnÞ ð5Þ

r ¼ �33:4F4:3aþ 57:1F4:1; R2

¼ 0:76 ðEtesian windsÞ ð6Þ

r ¼ �36:5F4:9aþ 62:9F4:5 R2

¼ 0:73 ðsea breezeÞ ð7Þ

From the above it is clear that the correlation betweenα and particle size is negative, as expected, since lowerα values are associated with larger particles. Thecoefficient of determination, R2, is high in all cases,except in winter. Nevertheless, in all cases the correla-tions are statistically significant at the 95% confidencelevel.

Fig. 4. Mean diurnal variation of particle radius in the atmosphere ofAthens during the warm period of 1997 (under strong Etesian winds:×; under fully developed sea breeze: ⋄).



3.3. Aerosol optical properties and back trajectories

Air mass back trajectories ending at Athens at threealtitudes 500, 1500 and 4000 m a.s.l. were calculated bythe HYSPLIT model (Draxler and Rolph, 2003). The re-analysis data from the National Center for Environmen-tal Prediction (NCEP) were used as input to the model.The trajectories end at Athens at 10 UTC (12:00 LST),which is the most representative hour regarding thespectral measurements of this study. Previous investiga-tions on Saharan dust outbreaks over Athens (Papayan-nis et al., 2005), and model analysis (Alpert et al., 2004)show that the dust outflow from Sahara to the Centraland Eastern Mediterranean occurs mainly in the first4–5 km. Therefore, the selection of the 4000-m altitudefor the identification of the Saharan dust particles is ap-propriate. The trajectory ending at 1500 m is associatedwith cases of dust transported in the lower atmosphericlayers, or when the mixed layer is low as in the winter.Conversely, when air masses originate from otherregions (e.g. continental Europe), they are generallyobserved within the lowest 2–3 km (Lelieveld et al.,2002). In the present study the trajectories at 4000 mprovided information on the transport of desert dust thatgenerally travels over the mixed layer. In addition, thetrajectories at 500 and 1500 m are indicative of thecirculation in the lower troposphere, mainly anthropo-

genic aerosols or vertical dust transport. Each computedtrajectory was associated with the corresponding aerosoloptical characteristics (AOD500, α320–575, r). However,relating column-integrated quantities to trajectories atspecific altitudes may be problematic and does not give aclear view of the dominant aerosol type. It was alsofound that the definition of the wind sector might bedifferent proportionally to the air mass altitude selected.Nevertheless, back trajectories are widely used in con-nection with AOD and α-Ångström (Pace et al., 2006) inorder to identify the aerosol types.

Due to the limited number of months (11) and days(42), the present study does not aim at conductinga climatological analysis of the aerosol properties inrelation to the air mass origin. Therefore, the air masstrajectories are used here to reveal some specific events.To this respect, no restricted criteria to the sector iden-tification were taken into account as in Pace et al. (2006)and Meloni et al. (in press). The air-mass origin can be(i) from the Saharan desert, (ii) from the Atlantic Ocean,(iii) from continental Europe and (iv) from local sources.It was found that 16 air mass types originated from thenorth (mainly in late summer) due to the Etesian winds.Eleven cases were associated with Saharan air masses(mainly in June and July) carrying desert-dust particles.Air masses having a clear Atlantic or marine originare very rare over Athens and are depicted in only two

Fig. 5. Four-day air mass trajectories at three levels on (a) 19/6/1997 (sea breeze circulation) and (b) 5/9/1997 (Etesian winds).



cases. During winter months, the air masses exhibited asimilar trajectory pattern before reaching Athens,originating from the North Atlantic and traversing theBoreal Sea and North Europe. In two cases the air masseswere characterized as local, enriched with anthropogenicaerosols, while in four cases their sector identificationwas very difficult (rather impossible) since air massesfrom different directions arrived over Athens at all threealtitudes.

In the present study, two characteristic situations ofdifferent wind regimes (Etesians and sea breeze), dif-ferent air mass trajectories (North and South directions)and different aerosol load and optical properties wereanalyzed. The first case refers to 19/6/1997, where thewind flow was from a southwestern direction with fulldevelopment of a sea breeze cell over Athens. Thesecond corresponds to 5/9/1997, a day with strongEtesianwinds. The 4-day air mass trajectories (Fig. 5a, b)from the HYSPLIT model clearly verify these condi-tions. On 19/6/1997 the air masses at all altitudes reachedAthens from the south having a clear Saharan origin,thus carrying desert-dust aerosols over the study area. In

contrast, on 5/9/1997 the air masses originated fromEastern Europe and former USSR countries carryingfine-mode particles from the combustion processes orforest fires from these regions. As clearly stated in Baliset al. (2003) this direction favors the presence of fine-mode aerosols over Northern Greece. Note here, that theair mass at 4000 m exhibits a descending route beforereaching Athens, while the air masses at the lower levelstraverse quite longer distances from those in the previouscase. This indicates that during those days the passage ofthe air masses above Eastern Europe was very fast,strongly associated with the intense Etesian winds. Thediurnal variation of the aerosol physical properties inthese two characteristic situations is shown in Fig. 6.

As far as the particle radius is concerned, this ishigher on 19/6/1997, the day with sea breeze circulationand the influence of African air masses. This fact wasexpected since the air masses from the Saharan desertcarry dust particles, which are usually large in size,r N0.6 μm (Dubovik et al., 2002). The mean diurnalparticle radius on that day is 38.04±15.31, while on theday with Etesian winds it was 28.89±13.20. This case

Fig. 6. Diurnal variation of (a) particle radius, (b) Ångström exponent α (320–575 nm), (c) aerosol optical depth at 350 nm and (d) aerosol opticaldepth at 500 nm on 19/6/1997 (×) and on 5/9/1997 (⋄).



study compares well with the mean situation presentedin Fig. 4. The coarse-mode dust aerosols have a cleareffect on the α values and dramatically reduce them on19/6/1997 (0.53±0514) compared to the day with north-sector air masses (0.85±0.42). It should be noted thatduring periods with winds from northerly directions theα values computed in Northern Greece were found to bemuch higher, directly influenced by the fine-mode aero-sols originating from the Balkan counties (Balis et al.,2003; Gerasopoulos et al., 2003). The lower α valuesover Athens are probably attributed to the mixingprocesses in the atmosphere and the enhanced presenceof marine aerosols during air-mass passage above theAegean Sea. The light southwesterly winds in conjunc-tion with the transport of desert-dust aerosols contributeto the enhanced aerosol load on 19/6/1997 compared tothe clearer day affected by the Etesian winds (5/9/1997).This fact is clearly depicted in the AOD field at both 350and 500 nm. Note that in the early morning hours theAOD field is similar for the 2 days, directly influencedby the lower mixing-layer height. Nevertheless, as theday progresses the two opposite wind regimes affectsignificantly the aerosol load and aerosol optical prop-erties. This fact is also evident in r and α320–575 values,where the differences increase after midday. Also, itshould be noted that the aerosol load dramatically de-creased after midday on the day with the Etesian windssince their flow strengthened in the early afternoon hours.Therefore, the main conclusion is that sea breeze and, ingeneral, southwesterly winds favor the accumulation ofaerosols over Athens, enhance its particle size due tocondensation, coagulation and humidification processesand reduce the spectral variation of the aerosol opticaldepth and Ångström exponent. Inversely, the influenceof Etesian winds constitutes a ventilation tool for theAthens atmosphere, removing the larger aerosol particlesleading to more transparent atmospheric conditions.

4. Concluding remarks

The study of the aerosol size in the atmosphere ofAthens revealed that small particles are abundant inwintertime, while larger particles are found in summer-time. The particle radius varies with respect to season,time and source of aerosols. The diurnal variationdepicts the changes in emission rates of aerosols and airmass advection and convection mechanisms.

The wind flow has a strong impact on atmosphericparticle size. The particles in the atmosphere of Athensappear to have greater radii under the influence of seabreeze cells compared to those under Etesian wind flow.The particle radius turns out to peak in the morning

during both flow conditions. This is attributed to thetraffic since it coincides with the rush hour of the cityand is favored by the appearance of a temperatureinversion. In the afternoon, smaller peaks may be due tophotochemical production and the generation of sec-ondary aerosol products. A general decreasing trend isobserved for both wind regimes.

It should be emphasized that the particle radiusresults derived in this work refer to all types of aerosolsincluded in a column originating at the site of themeasurements and ending at the top of the atmosphere inthe direction of the sun. Therefore, their size should beconsidered representative of the aerosol populationresident in the column during the time of the solarradiation measurements.

With the installation of a CIMEL sunphotometer atNOA during 2007 it is hoped that the knowledge of theaerosol-size distribution, its seasonal and diurnal varia-tions as well as its relations with the wind regimes, willgrow.

Acknowledgment

Dr. V. Cachorro and Dr. R. Vergaz (University ofValladolid, Spain) are gratefully acknowledged for pro-viding the authors with the algorithm LVEC for Mie-theory computations.

References

Abel, S.J., Highwood, E.J., Haywood, J.M., Stringer, M.A., 2005. Thedirect radiative effect of biomass burning aerosols over southernAfrica. Atmos. Chem. Phys. 5, 1999–2018.

Adamopoulos, A.D., Kambezidis, H.D., Zevgolis, D., Topalis, F.B.,2000. Estimation of total ozone column over Athens using ground-based beam solar irradiance measurements. Fres. Environ. Bull. 9,201–208.

Adamopoulos, A.D., Kambezidis, H.D., Zevgolis, D., 2005. Totalatmospheric transmittance in the UV and VIS spectra in Athens,Greece. Pure Appl. Geophys. 162, 409–431.

Alpert, P., Kishcha, P., Shtivelman, A., Krichak, S.O., Joseph, J.H.,2004. Vertical distribution of Saharan dust based on 2.5-yearmodel predictions. Atmos. Res. 70, 109–130.

Antoine, D., Nobileau, D., 2006. Recent increase of Saharan dusttransport over the Meditterenean Sea, as revealed from ocean colorsatellite (SeaWiFS) observations. J. Geophys. Res. 111, D12214.doi:10.1029/2005JD006795.

Asimakopoulos, D.N., Deligiorgi, D., Drakopoulos, C., Helmis, C.,Kokkori, K., Lalas, D., Sikiotis, D., Varotsos, C., 1992. Anexperimental study of nighttime air pollutant transport overcomplex terrain in Athens. Atmos. Environ. 26, 59–71.

Balis, D.S., Amiridis, V., Zerefos, C.S., Gerasopoulos, E., Andrae, M.,Zanis, P., Kazantzidis, A., Kazadzis, S., Papayannis, A., 2003.Raman lidar and sunphotometric measurements of aerosol opticalproperties over Thessaloniki, Greece during a biomass-burningepisode. Atmos. Environ. 37, 4529–4538.



Barnaba, F., Gobbi, G.P., 2004. Aerosol seasonal variability over theMediterranean region and relative impact of maritime, continentaland Saharan dust particles over the basin from MODIS data in theyear 2001. Atmos. Chem. Phys. 4, 2367–2391.

Bates, T.S., Huebert, B.J., Gras, J.L., Griffiths, F.B., Durkee, P.A.,1998. The international global atmospheric chemistry (IGAC)project's first aerosol characterization experiment (ACE 1):Overview. J. Geophys. Res. 103, 16297–16318.

Cachorro, V.E., 1998. Simple approaches and inversion methods toretrieve particle size parameters of atmospheric desert aerosols.Atmos. Environ. 32, 239–245.

Cachorro, V.E., de Frutos, A.M., 1997. An analytical study about theratio between particle mass loading and extinction: application todesert dust aerosols. J. Quant. Spectrosc. Radiat. Transfer 57,559–568.

Cachorro, V.E., Salcedo, L.L., 1991. New improvements for Miescattering calculations. J. Electromagn. Waves Appl. 15, 913–926.

Cachorro, V.E., Tanré, D., 1997. The correlation between particle massloading and extinction: application to desert dust aerosol contentestimation. Remote Sens. Environ. 60, 187–194.

Cachorro, V.E., Vergaz, R., de Frutos, A.M., 1999. A method foratmospheric aerosols retrieval based on extinction data: asensitivity analysis. 4th Conf. On Electromagnetic and LightScattering by non-spherical particles, Theory and Applications,Vigo, Spain.

Cachorro, V.E., Duran, P., Vergaz, R., de Frutos, A.M., 2000. Columnarphysical and radiative properties of atmospheric aerosols in northcentral Spain. J. Geophys. Res. 105 (D6), 7161–7175.

Carrico, C.M., Kus, P., Rood, M.J., Quinn, P.K., Bates, T.S., 2003.Mixtures of pollution, dust, sea salt, and volcanic aerosol duringACE-Asia: radiative properties as a function of relative humidity.J. Geophys. Res. 108 (D23), 8650. doi:10.1029/2003JD003405.

Charlson, R.J., Langner, J., Rodhe, H., Leovy, C.B., Warren, S.G.,1991. Perturbation of the northern hemisphere radiative balance bybackscattering from anthropogenic sulfate aerosols. Tellus 43,152–163.

Cheney, W., Kincaid, D., 1980. Numerical Mathematics andComputing. Brooks-Cole Publishing Comp., U.S.A.

Christides, A.I. 1995. A study on distribution of airborne pollutants inthe Thriassion Plain in Elefsina area, PhD Thesis, NationalTechnical University of Athens, Greece (in Greek).

Clappier, A., Martilli, A., Grossi, P., Thunis, P., Pasi, F., Krueger, B.,Calpini, B., Graziani, G., Van Den Bergh, H., 2000. Effect of seabreeze on air-pollution in the greater Athens area. Part I: numericalsimulations and field observations. J. Appl. Meteor. 39, 546–562.

D'Almeida, G.A., Koepke, P., Shettle, E.P., 1991. AtmosphericAerosol: Global Climatology and Radiative Characteristics. A.Deepack, Hampton, Va.

Day, D.E., Malm, W.C., 2001. Aerosol light scattering measurementsas a function of relative humidity: a comparison betweenmeasurements made at three different sites. Atmos. Environ. 35,5169–5176.

DEARTH (Directorate of Air Quality & Noise Control), 1998. TheAtmospheric Pollution inAthens—1997. Bulletin of theMinistry forthe Environment, Physical Planning and Public Works (in Greek).

Draxler, R.R., Rolph, G.D., 2003. HYSPLIT (Hybrid Single-ParticleLagrangian Integrated Trajectory) Model. NOAA Air ResourcesLaboratory, Silver, Spring, MD. http://www.arl.noaa.gov/ready/hysplit4.html.

Dubovik, O., King, M.D., 2000. A flexible inversion algorithm forretrieval of aerosol optical properties from sun and sky radiancemeasurements. J. Geophys. Res. 105, 20673–20696.

Dubovik, O., Holben, B., Eck, T.F., Smirnov, A., Kaufman, Y.J., King,M.D., Tanré, D., Slutsker, I., 2002. Variability of absorption andoptical properties of key aerosol types observed in worldwidelocations. J. Atmos. Sci. 59, 590–608.

Eleftheriadis, K., Balis, D., Ziomas, I.C., Colbeck, I., Manalis, N.,1998. Atmospheric aerosol and gaseous species in Athens, Greece.Atmos. Environ. 32, 2183–2191.

Ferrare, R., et al., 2000. Comparison of aerosol optical properties andwater vapor among ground and airborne lidars and sunphotometers during TARFOX. J. Geophys. Res. 105, 9917–9933.

Gerasopoulos, E., Andreae, M.O., Zerefos, C.S., Andreae, T.W., Balis,D., Formenti, P., Merlet, P., Amiridis, V., Parastefanou, C., 2003.Climatological aspects of aerosol optical properties in NorthernGreece. Atmos. Chem. Phys. 3, 2025–2041.

Giavis, G.M., Kambezidis, H.D., Sifakis, N., Tóth, Z., Adamopoulos,A.D., Zevgolis, D., 2005. Diurnal variation of the aerosol opticaldepth for two distinct cases in the Athens area, Greece. Atmos.Res. 78, 79–92.

Hänel, G., 1976. The properties of atmospheric aerosol particles asfunctions of relative humidity at thermodynamic equilibrium withsurrounding moist air. Advances in Geoph, vol. 19. AcademicPress, New York, pp. 73–188.

Haywood, J.M., Boucher, O., 2000. Estimates of the direct and indirectradiative forcing due to tropospheric aerosols: a review. Rev.Geophys. 38, 513–543.

Horvath, H., 1996. Spectral extinction coefficients of rural aerosol insouthern Italy—a case study of cause and effect of variability ofatmospheric aerosol. J. Aerosol Sci. 27, 437–453.

Horvath, H., Arboledas, L.A., Olmo, F.J., Jovanovic, O., Gangl, M.,Kaller, W., Sanchez, C., Sauerzopf, H., Seidl, S., 2002. Opticalcharacteristics of the aerosol in Spain and Austria and its effect onradiative forcing. J. Geophys. Res. 107 (D19), 4386.

Kambezidis, H.D., Tulleken, R., Amanatidis, G.T., Paliatsos, A.G.,Asimakopoulos, D.N., 1995. Statistical evaluation of selected airpollutants in Athens, Greece. Environmetrics 6, 349–361.

Kambezidis, H.D., Petrova, V.D., Adamopoulos, A.D., 1997a.Radiative transfer. I. Atmospheric transmission monitoring withmodeling and ground-based multispectral measurements. Appl.Opt. 36, 6976–6982.

Kambezidis, H.D., Petrova, V.D., Adamopoulos, A.D., 1997b. Radiativetransfer. II. Impact of meteorological variables and surface albedo onatmospheric optical properties retrieved from ground-based multi-spectral measurements. Appl. Opt. 36, 6983–6988.

Kambezidis, H.D., Weidauer, D., Melas, D., Ulbricht, M., 1998a. Airquality in the Athens basin during sea-breeze and non-sea-breezedays using laser-remote-sensing technique. Atmos. Environ. 32(12), 2173–2182.

Kambezidis, H.D., Katevatis, E.M., Petrakis, M., Lykoudis, S.,Asimakopoulos, D.N., 1998b. Estimation of the Linke andUnsworth–Monteith turbidity factors in the visible spectrum:application for Athens, Greece. Solar Energy 62, 39–50.

Kambezidis, H.D., Adamopoulos, A.D., Zevgolis, D., 2000a. Casestudies of spectral atmospheric transmittance in the ultraviolet andvisible regions in Athens, Greece. I: total transmittance. Atmos.Res. 54, 223–232.

Kambezidis, H.D., Adamopoulos, A.D., Zevgolis, D., 2000b. Casestudies of spectral atmospheric transmittance in the ultraviolet andvisible regions in Athens, Greece. II: aerosol transmittance. Atmos.Res. 54, 233–243.

Kambezidis, H.D., Adamopoulos, A.D., Gueymard, C., 2001a. TotalNO2 column amount over Athens, Greece in 1996–97. Atmos.Res. 57, 1–8.



Kambezidis, H.D., Adamopoulos, A.D., Zevgolis, D., 2001b.Determination of Ångström and Schüepp's parameters fromground-based spectral measurements of beam irradiance in theultraviolet and visible spectrum in Athens, Greece. Pure Appl.Geophys. 158, 821–838.

Kambezidis, H.D., Adamopoulos, A.D., Zevgolis, D., 2005. Spectralaerosol transmittance in the ultraviolet and visible spectrum inAthens, Greece. Pure Appl. Geophys. 162, 625–647.

Kaskaoutis, D.G., Kambezidis, H.D., 2006. Investigation on thewavelength dependence of the aerosol optical depth in the Athensarea. Q. J. R. Meteorol. Soc. 132, 2217–2235.

Kaskaoutis, D.G., Kambezidis, H.D., Adamopoulos, A.D., Kassome-nos, P.A., 2006a. Comparison between experimental data andmodeling estimates of aerosol optical depth over Athens, Greece.J. Atmos. Sol. Terr. Phys. 68, 1167–1178.

Kaskaoutis, D.G., Kambezidis, H.D., Adamopoulos, A.D., Kassome-nos, P.A., 2006b. On the determination of aerosols using theÅngström exponent in the Athens area. J. Atmos. Sol. Terr. Phys.68, 2147–2143.

King, M.D., Byrne, D.M., Herman, B.M., Reagan, J.A., 1978. Aerosolsize distributions obtained by inversion of spectral optical depthmeasurements. J. Atmos. Sci. 35, 2153–2167.

Kinne, S., Pueschel, R., 2001. Aerosol radiative forcing for Asiancontinental outflow. Atmos. Environ. 35, 5019–5028.

Klemm, O., Ziomas, I.C., Balis, D., Suppan, P., Slemr, J., Romero, R.,Vyras, L., 1998. A summer air-pollution study in Athens, Greece.Atmos. Environ. 32, 2071–2087.

Lelieveld, J., et al., 2002. Global air pollution crossroads over theMediterranean. Science 298, 794–999.

Markowicz, K.M., Flatau, P.J., Ramana, M.V., Crutzen, P.J.,Ramanathan, V., 2002. Absorbing Mediterranean aerosols lead toa large reduction in the solar radiation at the surface. Geophys. Res.Lett. 29. doi:10.1029/2002GL015767.

Meloni, D., di Sarra, A., Biavati, G., DeLuisi, J.J., Monteleone, F.,Pace, G., Piacentino, S., Sferlazzo, D.M., in press. Seasonalbehavior of Saharan dust events at the Mediterranean island ofLambedusa in the period 1999–2005. Atmos. Env. doi:10.1016/j.atmosenv.2006.12.001.

Molnàr, A., Mészáros, E., 2001. On the relation between the size andchemical composition of aerosol particles and their opticalproperties. Atmos. Environ. 35, 5053–5058.

Pace, G., di Sarra, A., Meloni, D., Piacentino, S., Chamard, P., 2006.Aerosol optical properties at Lambeduca (Cenral Mediterranean).1. Influence of transport and identification of different aerosoltypes. Atmos. Chem. Phys. 6, 697–713.

Papayannis, A., Balis, D., Amiridis, V., Chourdakis, G., Tsaknakis, G.,Zerefos, C.S., Castanho, A.D.A., Nickovic, S., Kazadzis, S.,

Grabowski, J., 2005. Measurements of Saharan dust aerosols overthe Eastern Mediterranean using elastic backscatter-Raman lidar,spectrophotometric and satellite observations in the frame of theEARLINET project. Atmos. Chem. Phys. 5, 2065–2079.

Reid, J.S., Eck, T.F., Christopher, S.A., Hobbs, P.V., Holben, B.N.,1999. Use of the Ångström exponent to estimate the variability ofoptical and physical properties of aging smoke particles in Brazil.J. Geophys. Res. 104 (D22), 27473–27489.

Scheff, P.A., Valiozis, C., 1990. Characterization and sourceidentification of respirable particulate matter in Athens, Greece.Atmos. Environ. 24, 203–211.

Schmid, B., Matzler, C., Heimo, A., Kampfer, N., 1997. Retrieval ofoptical depth and particle size distribution of tropospheric andstratospheric aerosols by means of sun photometry. IEEE Trans.Geosci. Remote Sens. 35, 172–182.

Schuster, G.L., Dubovik, O., Holben, B.N., 2006. Ångström exponentand bimodal aerosol size distributions. J. Geophys. Res. 111,D07207. doi:10.1029/2005JD006328.

Seinfeld, J.H., Pandis, S.N., 1997. Atmospheric Chemistry andPhysics: From Air Pollution to Climate Change. John Wiley,New York, p. 1326.

Smirnov, A., Holben, B.N., Kaufman, Y.J., Dubovic, O., Eck, T.F.,Slutsker, I., Pietras, C., Halthore, R.N., 2002. Optical properties ofatmospheric aerosol in maritime environments. J. Atmos. Sci. 59,501–523.

Suppan, P., Fabian, P., Vyras, L., Gryning, S.E., 1998. The behaviourof ozone and peroxyacetyl nitrate concentrations for different windregimes during the MEDCAPHOT-TRACE campaign in thegreater Athens area of Athens, Greece. Atmos. Environ. 32,2089–2102.

Vergaz, R., Cachorro, V.E., Duran, P., de Frutos, A.M., Vilaplana,J.M., de la Morena, B., 1999. Monitoring turbidity parameters andradiative properties of atmospheric aerosols in two differentSpanish climatic sites. 4th Conf. On Electromagnetic and LightScattering by non-spherical particles, Theory and Applications,Vigo, Spain.

Vitale, V., Tomasi, C., Lupi, A., Cacciari, A., Marani, S., 2000.Retrieval of columnar aerosol size distributions and radiative-forcing evaluations from sun-photometric measurements takenduring the CLEARCOLUMN (ACE 2) experiment. Atmos.Environ. 34, 5095–5105.

Yaaqub, R.R. 1989. Composition of atmospheric aerosols in rural ofEast Anglia and meteorological controls, PhD Thesis, Universityof East Anglia, U.K.

Yamamoto, G., Tanaka, M., 1969. Determination of aerosol sizedistribution from spectral attenuation measurements. Appl. Opt. 8,447–453.


a study of aerosol particle sizes in the atmosphere of athens, greece, retrieved from solar spectral...

Documents