The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data

The paper presents the results of the monthly mean PM₂.₅ analysis in the period from 2005 to 2013 over the Europe based on the connection between daily fine particle concentrations (PM₂.₅) by surface in-situ measurements in AIRBASE network and column aerosol optical thickness (AOT) derived from POLD...

Повний опис

Збережено в:
Бібліографічні деталі
Опубліковано в: :Advances in Astronomy and Space Physics
Дата:2013
Автор: Bovchaliuk, A.
Формат: Стаття
Мова:English
Опубліковано: Головна астрономічна обсерваторія НАН України 2013
Онлайн доступ:https://nasplib.isofts.kiev.ua/handle/123456789/119625
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data / A. Bovchaliuk // Advances in Astronomy and Space Physics. — 2013. — Т. 3., вип. 2. — С. 102-108. — Бібліогр.: 27 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-119625
record_format dspace
spelling Bovchaliuk, A.
2017-06-07T18:21:23Z
2017-06-07T18:21:23Z
2013
The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data / A. Bovchaliuk // Advances in Astronomy and Space Physics. — 2013. — Т. 3., вип. 2. — С. 102-108. — Бібліогр.: 27 назв. — англ.
2227-1481
https://nasplib.isofts.kiev.ua/handle/123456789/119625
The paper presents the results of the monthly mean PM₂.₅ analysis in the period from 2005 to 2013 over the Europe based on the connection between daily fine particle concentrations (PM₂.₅) by surface in-situ measurements in AIRBASE network and column aerosol optical thickness (AOT) derived from POLDER-3/PARASOL satellite sensor. The regression function between PM₂.₅ and AOT was derived from measurements done over Europe in the period from April to October 2007. Considering 749 match-up data points over 20 fine particle monitoring sites, we found that the POLDER-3/PARASOL derived AOT at 865 nm is correlated with collocated PM₂.₅ measurements with a correlation coeficient 0.62 (RMS = 3.26). According to the obtained linear regression PM2.5 = 73.4 × AOT865 + 9.6, a signi cant o set caused an introduction of the threshold of 0.01 in monthly mean AOT for assessment of PM₂.₅ based on satellite data. Therefore, only PM₂.₅ values larger than 10.3 µg/m³ can be obtained using this method. According to results the monthly mean PM₂.₅ in the period from 2005 to 2013 over the Europe is usually characterised by values less than 12 µg/m³ (classified as "good" by Air Quality Categories, AQC), but values ranging from 12 to 18 µg/m³ (classified as "moderate") are found in the densely populated and industrial areas, such as the Netherlands, Belgium, the Ruhr and Danube area, Northern Italy, Poland, Romania and Eastern Ukraine. Additionally, the maximum values of PM₂.₅ over Eastern Europe are observed during forest, peat and agricultural wildfires in May 2006 (15-21 µg/m³), April 2009 (14-18 µg/m³) and August 2010 (35-55 µg/m³, classi-fied as "unhealthy for sensitive groups"). An extended set of aerosol parameters including particle size distribution, complex refractive index, as well as parameters characterising aerosol particle shape and vertical distribution will be analysed in the future work.
This study was supported by Award No. UKG2-2969-KV-09 of the U.S. Civilian Research & Development Foundation (CRDF) and by the projects M/115-2012, and F41/106-2012 of Derzhinformnauka, and PICS 2013-2015 of CNRS and NASU, and the project 11BF051-01-12 of the Taras Shevchenko National University of Kyiv. The author thanks the ICARE Data and Services Center team for providing access to the POLDER-3 data and for general assistance and development support. The author acknowledges all network operators that submitted their data to the AIRBASE database. The author is grateful to G. Milinevsky, V. Danylevsky, V. Bovchaliuk, P. Goloub, O. Dubovik, F. Ducos, M. Sosonkin and anonymous referee for detailed comments and useful suggestions.
en
Головна астрономічна обсерваторія НАН України
Advances in Astronomy and Space Physics
The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data
spellingShingle The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data
Bovchaliuk, A.
title_short The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data
title_full The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data
title_fullStr The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data
title_full_unstemmed The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data
title_sort spatial variability of pm₂.₅ over europe using satellite polder-3/parasol data
author Bovchaliuk, A.
author_facet Bovchaliuk, A.
publishDate 2013
language English
container_title Advances in Astronomy and Space Physics
publisher Головна астрономічна обсерваторія НАН України
format Article
description The paper presents the results of the monthly mean PM₂.₅ analysis in the period from 2005 to 2013 over the Europe based on the connection between daily fine particle concentrations (PM₂.₅) by surface in-situ measurements in AIRBASE network and column aerosol optical thickness (AOT) derived from POLDER-3/PARASOL satellite sensor. The regression function between PM₂.₅ and AOT was derived from measurements done over Europe in the period from April to October 2007. Considering 749 match-up data points over 20 fine particle monitoring sites, we found that the POLDER-3/PARASOL derived AOT at 865 nm is correlated with collocated PM₂.₅ measurements with a correlation coeficient 0.62 (RMS = 3.26). According to the obtained linear regression PM2.5 = 73.4 × AOT865 + 9.6, a signi cant o set caused an introduction of the threshold of 0.01 in monthly mean AOT for assessment of PM₂.₅ based on satellite data. Therefore, only PM₂.₅ values larger than 10.3 µg/m³ can be obtained using this method. According to results the monthly mean PM₂.₅ in the period from 2005 to 2013 over the Europe is usually characterised by values less than 12 µg/m³ (classified as "good" by Air Quality Categories, AQC), but values ranging from 12 to 18 µg/m³ (classified as "moderate") are found in the densely populated and industrial areas, such as the Netherlands, Belgium, the Ruhr and Danube area, Northern Italy, Poland, Romania and Eastern Ukraine. Additionally, the maximum values of PM₂.₅ over Eastern Europe are observed during forest, peat and agricultural wildfires in May 2006 (15-21 µg/m³), April 2009 (14-18 µg/m³) and August 2010 (35-55 µg/m³, classi-fied as "unhealthy for sensitive groups"). An extended set of aerosol parameters including particle size distribution, complex refractive index, as well as parameters characterising aerosol particle shape and vertical distribution will be analysed in the future work.
issn 2227-1481
url https://nasplib.isofts.kiev.ua/handle/123456789/119625
citation_txt The spatial variability of PM₂.₅ over Europe using satellite POLDER-3/PARASOL data / A. Bovchaliuk // Advances in Astronomy and Space Physics. — 2013. — Т. 3., вип. 2. — С. 102-108. — Бібліогр.: 27 назв. — англ.
work_keys_str_mv AT bovchaliuka thespatialvariabilityofpm25overeuropeusingsatellitepolder3parasoldata
AT bovchaliuka spatialvariabilityofpm25overeuropeusingsatellitepolder3parasoldata
first_indexed 2025-11-25T15:45:25Z
last_indexed 2025-11-25T15:45:25Z
_version_ 1850517124802412544
fulltext The spatial variability of PM2.5 over Europe using satellite POLDER-3/PARASOL data A.Bovchaliuk ∗ Advances in Astronomy and Space Physics, 3, 102-108 (2013) © A.Bovchaliuk, 2013 Space Physics Laboratory, Astronomy and Space Physics Department, Taras Shevchenko National University of Kyiv, Volodymyrska str. 64/13, 01601 Kyiv, Ukraine The paper presents the results of the monthly mean PM2.5 analysis in the period from 2005 to 2013 over the Europe based on the connection between daily �ne particle concentrations (PM2.5) by surface in-situ measure- ments in AIRBASE network and column aerosol optical thickness (AOT) derived from POLDER-3/PARASOL satellite sensor. The regression function between PM2.5 and AOT was derived from measurements done over Europe in the period from April to October 2007. Considering 749 match-up data points over 20 �ne particle monitoring sites, we found that the POLDER-3/PARASOL derived AOT at 865 nm is correlated with collocated PM2.5 measurements with a correlation coe�cient 0.62 (RMS = 3.26). According to the obtained linear regression PM2.5 = 73.4×AOT865 + 9.6, a signi�cant o�set caused an introduction of the threshold of 0.01 in monthly mean AOT for assessment of PM2.5 based on satellite data. Therefore, only PM2.5 values larger than 10.3µg/m3 can be obtained using this method. According to results the monthly mean PM2.5 in the period from 2005 to 2013 over the Europe is usually characterised by values less than 12µg/m3 (classi�ed as �good� by Air Quality Categories, AQC), but values ranging from 12 to 18µg/m3 (classi�ed as �moderate�) are found in the densely populated and industrial areas, such as the Netherlands, Belgium, the Ruhr and Danube area, Northern Italy, Poland, Romania and Eastern Ukraine. Additionally, the maximum values of PM2.5 over Eastern Europe are observed during forest, peat and agricultural wild�res in May 2006 (15�21µg/m3), April 2009 (14�18µg/m3) and August 2010 (35�55µg/m3, classi- �ed as �unhealthy for sensitive groups�). An extended set of aerosol parameters including particle size distribution, complex refractive index, as well as parameters characterising aerosol particle shape and vertical distribution will be analysed in the future work. Key words: aerosols, PM2.5, atmosphere, AOT, pollution introduction Atmospheric aerosols are conventionally de�ned as particles suspended in air having radius in the range from 0.001µm to 100µm [7]. These particles have a signi�cant negative e�ects on human health [23], including lung cancer, asthmatic symptom, pul- monary in�ammation and cardiopulmonary mortal- ity. The solution of many practical problems in the �eld of environmental and human health protection is connected with needs for high quality information about pollution level and the physicochemical prop- erties of atmospheric aerosols. The particulate matter (PM) mass, also known as air pollution, is a generally accepted means to quantify the amount of aerosols in the atmosphere and is used as a standard to evaluate air quality by United States Environmental Protection Agency1 (EPA). Fine fraction of particulate matter (PM2.5, µg/m3) is the integrated mass of aerosols with di- ameter up to 2.5µm that is generally resulted from combustion, including motor vehicles, power plants, forest wild�res, and agricultural burning. The EPA has established that the 24-hour averaged PM2.5 con- centration must be less than 35.4µg/m3 for healthy conditions concerning the whole population. Most of monitoring stations are located close to major ur- ban regions leaving large areas without operational observations. Furthermore, it is widely recognised that it is problematic to measure the absolute level of PM2.5 on a routine basis. In routine measure- ment processes, heating of the air sample is neces- sary, which partially increase volatilisation of semi- volatile aerosol components. This in turn leads to systematic measurement errors depending on mea- surement technique and aerosol composition. The limited spatial representativeness and di�erent sys- tematic errors of national or regional air quality net- works make it impossible to achieve an air quality overview across Europe by surface in-situ measure- ments only. In addition to surface in-situ measurements of PM2.5, the monitoring of aerosols and trace gases from space has been widely used [16, 20, 26]. Pri- mary aerosol quantity derived from spaceborne re- mote sensors operated in the visual and IR wave- ∗bovchaliuk@gmail.com 1http://www.epa.gov 102 Advances in Astronomy and Space Physics A.Bovchaliuk length bands is the aerosol optical thickness (AOT). The relationship between column AOT derived from satellite and PM2.5 surface in-situ measurements has already been explored over the United States [1, 27], Italy [8] and some populated and indus- trial regions in Asia [18] using retrievals from the MODIS radiometer (MODerate Imaging Spectrora- diometer), and over France [15] based on data from the POLDER-2 (Polarization and Directionality of Earth's Re�ectances) satellite instrument. Currently, the comparison studies between new satellite radiometer POLDER-3 and PM2.5 surface in-situ measurements over Europe have not been con- ducted yet. In the present paper we use the relation obtained for 2007, being the year when PM2.5 data were available online. We present results on the spa- tial distributions of monthly mean PM2.5 values es- timations based on POLDER-3 data over Europe. Last section gives an overview of special events and PM2.5 characteristics in the period from 2005 to 2013 over investigated area. surface in-situ airbase data In this paper, the PM2.5 data from the AIR- BASE database2 were used for determine the rela- tionship between PM2.5 values and AOT retrieved from POLDER-3. AIRBASE is a public database system operated by the European Environmental Agency (EEA). It provides air quality monitoring data and associated information submitted by more than 30 participating countries within Europe [9]. Measurement sites are classi�ed according to the type of surroundings (urban, suburban, or rural) and dominant local source (tra�c, industrial, or back- ground). The reported precision of the measure- ments is satisfactory (random errors typically less than 15% for daily averages), and the absolute sys- tematic uncertainty in these data is estimated to be 19%. The gravimetric methods based on the deter- mination of the PM2.5 mass fraction of particulate matter collected on a �lter under ambient conditions was used to give typically 24-hour averages, while au- tomatic instrumental methods TEOM (Tapered Ele- ment Oscillating Microbalance) and Beta Absorption provided hourly data [21]. In the Beta Absorption and TEOM methods, the sampled air is heated be- fore collecting the aerosols on a �lter to avoid hu- midi�cation of the �lter. This procedure is known to lead to loss of semi-volatile aerosol components. The extent to which volatilisation occurs depends on time and space dependent chemical composition of the aerosols, and the measurement technique. To correct for systematic errors, correction factors are generally applied to the measured data, but these di�er by 30% or more between countries [6]. Conse- quently, systematic di�erences between territories of measurements may be present in the data due to use of di�erent methods. However, the data from AIR- BASE were used in the form that submitted by the countries. In the comparison of spatial distribution of AOT and PM2.5, 17 rural and 3 suburban back- ground stations have been analysed. The stations were selected in Austria, Belgium, Cyprus, Czech Republic, Finland, France, Germany, Italy, Latvia, Netherlands, Slovenia, Spain, Sweden, Switzerland and the United Kingdom to present di�erent under- lying surfaces (need for satellite retrieval) and aerosol loading. According to standard data from stations, the 24-hour averaged PM2.5 values were used in anal- ysis. satellite polder-3/parasol data A unique set of observations on the global aerosol distribution has been acquired by the POLDER- 3 instrument that was installed aboard French mi- crosatellite PARASOL in December 2004 and col- lected data in the period from March 2005 to September 2013. The instrument carried out multi- spectral (443, 490, 565, 670, 763, 765, 865, 910 and 1020 nm), multi-directional (as many as 16 direc- tions within the scope of 100◦ approximately along the ground trace) measurements of intensity and lin- ear polarization degree of the back-scattered solar radiation [25]. The current standard aerosol inver- sion strategy detailed in [10] is based on the look-up tables approach, where the re�ected radiances are simulated for 10 aerosol models with log-normal size distributions of particles with e�ective radius from 0.075 to 0.225µm and a complex refractive index of 1.47�0.01i. This inversion can be used to character- ize the �ne mode aerosols with particle size < 0.3µm [2, 5, 11]. According to the cloud-screening algorithm [4] the cloud-free pixels were processed only. The surface contribution to the polarized re�ectance was based on a priori values (as a function of observation geometry and surface type) derived from statistical analysis of POLDER-1 data [22]. The aerosol param- eters were adjusted to give the best agreement be- tween the measured and simulated multidirectional polarized radiances at 490, 670 and 865 nm wave- lengths. The spectroradiometer POLDER-3 data were processed and described at the project web-page3. The spatial resolution of POLDER-3 instrument was 16 km× 18 km in the AOT. The aerosol characteris- tics accuracy has been estimated over ocean [12, 13] and land [5, 11], where correlation with AERONET (AErosol RObotic NETwork) [14] data is equal to 0.77�0.95 depending on location of measurement and aerosol type. For example, the correlation coe�- cients for the POLDER-3 and AERONET AOT re- trieval comparisons are equal: 0.78 for Moscow site, 2http://acm.eionet.europa.eu/databases/airbase/ 3http://www.icare.univ-lille1.fr/parasol 103 Advances in Astronomy and Space Physics A.Bovchaliuk 0.76 � Minsk, 0.86 � Belsk, 0.81 � Moldova, 0.93 � Kyiv and 0.63 for Sevastopol sites [2]. comparison between aot and pm2.5 The satellite overpass time for the investigated territory was 10:00�11:30UT depending on the orbit and geometry of measurements. The POLDER-3 de- rived AOT values at 865 nm were selected from each pixel (16 km× 18 km) covered the PM2.5 surface in- situ stations. It should be noted that the amount of data increased for the sites at lower latitudes and larger number of sunshine days. For latitudes above 50◦N, data were not available in winter, because the solar elevation is too low at these latitudes to allow AOT retrieval. The underlying surface in the inves- tigated territory varies highly in spring and autumn due to variable vegetative cover. Moreover, the sur- face impact on polarization is relatively greater when the amount of aerosol in the atmosphere is small. Fig. 1: Regression line between POLDER-3/PARASOL derived AOT at 865 nm and PM2.5 measured at the sur- face over 20 stations in Europe from April to October 2007. For deriving the regression function PM2.5 ver- sus AOT the 24-hour averaged PM2.5 values were selected from 20 stations in Europe with more than 75% of valid measurements in the period from April to October 2007. Than we excluded AOT zero val- ues obtained by POLDER-3 (lack of phase angles, glint or impossibility of retrievals) from analysis as well. Taking into account the above requirements the comparison was performed between POLDER-3 de- rived AOT and PM2.5 values and presented in Fig. 1. The overall match-up database corresponds to 749 comparison points. A signi�cant o�set 9.6µg/m3 in- dicates that the satellite has a limited capacity for monitoring small amounts of aerosols. The slope of the regression contributes to 73.4µg/m3 per AOT unit (RMS = 3.26). The correlation coe�cient is es- timated as 0.62. The deviation could be explained by the rough spatial resolution of the POLDER-3 instrument comparing to the local station. Besides, the relationship between PM2.5 and AOT strongly depends on vertical pro�le and aerosols type with similar optical properties that will be evaluated in the future study. The comparison of relation between POLDER-3 AOT and PM2.5 with those previously established using MODIS is complicated by the in�uence of sev- eral factors, like using PM10 instead of PM2.5 [8, 18], di�erent wavelengths and spatial resolutions for the satellite AOT [27], di�erent time periods, and di�er- ent sensitivity of the sensors to �ne particles [5]. For example, the local study over Alabama [27] demon- strates that the slope is equal to 71µg/m3 per AOT unit using PM2.5 based on MODIS data that is in a good agreement with the results presented here. Since the PM2.5 values lie in the range from 3.2 to 33.8µg/m3, it can be argued that no speci�c events occurred in the days of comparison. PM2.5 values larger than 35.4µg/m3 are not presented, which characterise this match-up data by two �rst categories according to the EPA Air Quality Cate- gories (AQC) [27], i. e. below 12µg/m3 (classi�ed as �good�) and between 12.1µg/m3 and 35.4µg/m3 (classi�ed as �moderate�). Besides, the AOT at 870 nm over Europe during 2003�2011 years was characterized by values ranging from 0.002 to 0.2, except the period of July�August 2010 with strong forest and peat wild�res when the AOT typical val- ues ranged from 0.3 to 0.5 [2]. spatial distributions of monthly mean pm2.5 Based on analysis presented in previous section, it is possible to estimate the spatial distributions of monthly mean PM2.5 using POLDER-3/PARASOL AOT data during 2005�2013 over Europe. Never- theless, a signi�cant o�set 9.6µg/m3 indicates that atmosphere with small amount of aerosols cannot be represented on the maps. Thus, in order to decrease the uncertainties in determining PM2.5 the threshold of 0.01 was applied for monthly mean AOT values. This threshold was chosen due to increasing uncer- tainties in determining AOT when the amount of aerosol in the atmosphere is small. According to lin- ear regression (Fig. 1) only PM2.5 values larger than 10.3µg/m3 can be obtained using this method. The monthly mean PM2.5 values for August month through 2005�2013 except 2010 are presented in Fig. 2. The territory is usually characterised by PM2.5 values less 12µg/m3. In general, there are some changes in distribution of aerosol pollution in the same month from year to year due to regional emissions, variable transboundary transfers and dif- 104 Advances in Astronomy and Space Physics A.Bovchaliuk ferent quantity of sunshine days for averaging. In addition, areas with moderate PM2.5 values (12� 18µg/m3) were found in the densely populated and industrial regions of Europe. It should be noted that analysis of aerosol pollution over these areas was based on POLDER-3 data retrieved during the period from April 2005 to September 2013. According to [24] the MODIS instrument ob- served emissions from agricultural �res in the Baltic countries, Belarus, Ukraine and Russia in April and May 2006. The PM2.5 shows the high values ranging in 15�21µg/m3 in May 2006 over Belarus, North- ern Ukraine and west of the European part of Rus- sia seen in Fig. 3a, which was caused by biomass burning particles appeared during this event. More- over, values 18�21µg/m3 were observed over Fin- land which can be explained by transportation of these particles from wild�res that caused the most severe air pollution episode over the recorded period at research stations in Svalbard [19]. Furthermore, the aerosol transboundary transportation from for- est and brushwood �res in the European part of Rus- sia in April 2009 produced increased PM2.5 values of 14�18µg/m3 over Ukraine, Romania, Moldova and Eastern Belarus (Fig. 3b). As one can seen from Fig. 3a and Fig. 3b the PM2.5 values in the range of 17�19µg/m3 were observed over Belgium, Nether- lands, and Germany Ruhr area, that equally char- acterizes these industrial area in the spring of 2006 and 2009. The maximum PM2.5 values were revealed in August 2010 (Fig. 3c) that corresponds to wild- �res in the centre of the European part of Rus- sia [17]. Monthly mean values are ranged from 35 to 55µg/m3 over Moscow, Nizhniy Novgorod, Ryazan, Tula and Vladimir regions, and accord- ing to the AQC criteria this period can be classi- �ed as �unhealthy for sensitive groups�. Moreover, the biomass burning aerosols transboundary trans- portation to Ukraine [3], Belarus, Lithuania, Latvia and Estonia was observed with PM2.5 values from 20 to 35µg/m3 (Fig. 3c). For comparison to the presented events the monthly average PM2.5 val- ues for April 2013 are given in Fig. 3d, which are probably characterised by presence of anthropogenic aerosols over all Europe, particularly over East Ger- many (17�19µg/m3), Northern Italy (15�18µg/m3), Central and Northern France (13�16µg/m3), Cen- tral Ukraine and Moldova (12�15µg/m3). The short time events (hours, days) were not presented in this study due to monthly averaging data. conclusions The AOT values retrieved from POLDER-3 data during 2005�2013 over Europe have been used to characterize aerosol pollution near the surface. POLDER-3 AOT retrieval algorithm over land is based on the measurement of the linear polariza- tion of the light backscattered to space. It should be emphasised that standard algorithm is particu- larly sensitive to �ne mode aerosols that polarize the signal. The relationship between column aerosol optical thickness derived from satellite POLDER- 3/PARASOL and observed PM2.5 by surface in-situ measurements is not straightforward and depends on the vertical distribution of the particles and their optical and microphysical properties. The uncer- tainties of both types of measurements should be taken into consideration as well. In the case of the satellite inversion algorithm the major in�u- ence of the surface and the spatial resolution pro- duce those uncertainties. However, the POLDER- 3/PARASOL derived AOT at 865 nm is fairly well correlated with collocated daily PM2.5 measurements (PM2.5 = 73.4×AOT865 + 9.6). The correlation co- e�cient 0.62 and slope of the regression 73.4µg/m3 per AOT unit (RMS = 3.26) are derived based on AIRBASE data from 20 stations in Europe consid- ered in the present study. Consequently, a signi�cant o�set caused an introduction of the threshold of 0.01 in monthly mean POLDER-3/PARASOL AOT for assessment of PM2.5 based on satellite data. With the aim to provide possible explanations for the general agreement and di�erences between aerosol properties retrieved from satellites and those observed at surface in-situ stations, in future work we will analyse an extended set of aerosol parameters in- cluding particle size distribution, complex refractive index, as well as parameters characterizing aerosol particle shape and vertical distribution. It is planned to use GEOS-Chem model for retrieving the distri- bution of aerosol mass and AOT with a transport time step of 15min to calculate PM2.5 with better precision. The spatial distributions of monthly mean PM2.5 using POLDER-3/PARASOL AOT data have been �rstly estimated over Europe in 2005�2013. The Eu- rope territory is usually characterised by PM2.5 up to 12µg/m3, which is lower than in industrial regions of East Asia and India. The POLDER-3/PARASOL measurements show the major aerosol source regions in the Netherlands, Belgium, Germany, Northern Italy, Eastern Poland and Great Britain, Romania, Hungary, Moldova, Eastern Ukraine and Belarus, as well as individual large cities and industrial valleys (Moscow, London, Rhone, and Danube). The PM2.5 values are ranged from 12 to 18µg/m3 over these industrial regions in the period from April 2005 to September 2013. In addition, the maximum values of PM2.5 (15�21µg/m 3) over Eastern Europe have been observed during forest, peat and agricultural wild- �res in May 2006 over Belarus, Finland and North- ern Ukraine, in April 2009 � 14�18µg/m3 over Euro- pean part of Russia, Ukraine, Romania, Moldova and Eastern Belarus, and in August 2010 � 35�55µg/m3 105 Advances in Astronomy and Space Physics A.Bovchaliuk over Moscow, Nizhniy Novgorod, Ryazan, Tula and Vladimir region in Russia. Aerosol amount during August 2010 is classi�ed as �unhealthy for sensitive groups� according to AQC criteria. acknowledgement This study was supported by Award No. UKG2- 2969-KV-09 of the U.S. Civilian Research & Devel- opment Foundation (CRDF) and by the projects M/115-2012, and F41/106-2012 of Derzhinform- nauka, and PICS 2013-2015 of CNRS and NASU, and the project 11BF051-01-12 of the Taras Shevchenko National University of Kyiv. The author thanks the ICARE Data and Services Center team for providing access to the POLDER-3 data and for general assistance and development support. The author acknowledges all network operators that sub- mitted their data to the AIRBASE database. The author is grateful to G.Milinevsky, V.Danylevsky, V.Bovchaliuk, P.Goloub, O.Dubovik, F.Ducos, M. Sosonkin and anonymous referee for detailed com- ments and useful suggestions. references [1] Al-Saadi J., Szykman J., PierceR.B. et al. 2005, Bull. Am. Meteorol. Soc., 86, 1249 [2] BovchaliukA., MilinevskyG., DanylevskyV. et al. 2013, Atmos. Chem. Phys., 13, 6587 [3] BovchaliukV., BovchaliukA., MilinevskyG. et al. 2013, Advances in Astronomy and Space Physics, 3, 46 [4] BréonF.-M. & Colzy S. 1999, J. Appl. Meteorol., 38, 777 [5] BréonF.-M., VermeulenA. & Descloitres J. 2011, Re- mote Sensing of Environment, 115, 3102 [6] BuijsmanE. & DeLeeuwF. 2004, ETC/ACC Technical Paper 2004/4, Bilthoven, NL, 21 [7] ChinM., KahnR.A. & Schwartz S. E. 2009, National Aeronautics and Space Administration, Washington, D.C., USA, 128 [8] ChuD.A., KaufmanY. J., ZibordiG. et al. 2003, J. Geo- phys. Res., 108, 4661 [9] DeLeeuwF. 2012, ETC/ACM Technical Paper 2012/4, Bilthoven, NL, 57 [10] Deuzé J. L., BréonF.M., DevauxC. et al. 2001, J. Geo- phys. Res., 106, 4913 [11] FanX.H., GoloubP., Deuzé J. L. et al. 2008, Remote Sensing of Environment, 112, 697 [12] GoloubP., TanréD., Deuzé J. L. et al. 1999, IEEE Trans- actions on Geoscience and Remote Sensing, 37, 1586 [13] HermanM., Deuzé J.-L., MarchandA., RogerB. & Lal- lart P. 2005, J. Geophys. Res.: Atmospheres, 110, D10S02 [14] HolbenB.N., EckT. F., Slutsker I. et al. 1998, Remote Sensing of Environment, 66, 1 [15] KacenelenbogenM., Léon J.-F., Chiapello I. & TanréD. 2006, Atmos. Chem. Phys., 6, 4843 [16] KingM.D., KaufmanY. J., TanréD. & NakajimaT. 1999, Bull. Am. Meteorol. Soc., 80, 2229 [17] Konovalov I. B., BeekmannM., Kuznetsova I. N., YurovaA. & ZvyagintsevA.M. 2011, Atmos. Chem. Phys., 11, 10031 [18] LiC., LauA.K.-H., Mao J. & ChuD.A. 2005, IEEE Transactions on Geoscience and Remote Sensing, 43, 2650 [19] Lund MyhreC., ToledanoC., MyhreG. et al. 2007, At- mos. Chem. Phys., 7, 5899 [20] MartinR.V. 2008, Atmospheric Environment, 42, 7823 [21] MolW. J.A. & van HooydonkP.R. 2012, ETC/ACM Technical Paper 2012/1, 58 [22] Nadal F., BreonF.-M. 1999, IEEE Transactions on Geo- science and Remote Sensing, 37, 1709 [23] Pope III C.A., BurnettR.T., ThunM. J. et al. 2002, J. Am. Med. Assoc., 287, 1132 [24] StohlA., BergT., Burkhart J. F. et al. 2007, Atmos. Chem. Phys., 7, 511 [25] TanréD., BréonF.M., Deuzé J. L. et al. 2011, Atmo- spheric Measurement Techniques, 4, 1383 [26] VijayaraghavanK., Snell H. E. & SeigneurC. 2008, Envi- ronmental Science & Technology, 42, 8187 [27] Wang J. & Christopher S.A. 2003, Geophys. Res. Lett., 30, 2095 106 Advances in Astronomy and Space Physics A.Bovchaliuk Fig. 2: Spatial distributions of monthly mean PM2.5 for August month through 2005�2013 except 2010. White: missing data or ocean. 107 Advances in Astronomy and Space Physics A.Bovchaliuk Fig. 3: Spatial distributions of monthly mean PM2.5 for: a) May 2006; b) April 2009; c) August 2010; d) April 2013. White: missing data or ocean. Note di�erent scale for August 2010. 108