Measurements of particulate matter within the framework of the European Monitoring and Evaluation Programme (EMEP)

The Science of the Total Environment 285 Ž2002. 209᎐235

Measurements of particulate matter within the framework of the European Monitoring and Evaluation Programme ž EMEP/ I. First results Mihalis LazaridisU , Arne Semb, Steinar Larssen, Anne-Gunn Hjellbrekke, Øystein Hov, Jan Erik Hanssen, Jan Schaug, Kjetil Tørseth Norwegian Institute for Air Research (NILU), Post Box 100, NO-2027 Kjeller, Norway Received 27 March 2001; accepted 14 June 2001

Abstract Particulate matter ŽPM. monitoring presents a new challenge to the transboundary air pollution strategies in Europe. Evidence for the role of long-range transport of particulate matter and its significant association with a wide range of adverse health effects has urged for the inclusion of particulate matter within the European Monitoring and Evaluation Programme ŽEMEP. framework. Here we review available data on PM physico-chemical characteristics within the EMEP framework. In addition we identify future research needs for the characterisation of the background PM in Europe that include detailed harmonised measurements of mass, size and chemical composition Žmass closure. of the ambient aerosol. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Particulate matter; Regional aerosol; Chemical characteristics

1. Introduction Long-range transport of gaseous air pollutants has been studied extensively in Europe during the last decades under the framework of the European Monitoring and Evaluation Program U

Corresponding author. Present address: Technical University of Crete, Department of Environmental Engineering, Chania, Greece. Fax: q30-821-37474. E-mail address: [email protected] ŽM. Lazaridis..

ŽEMEP. which is an acronym for the Co-operative Programme for Monitoring and Evaluation of the Long-range Transmission of air Pollutants in Europe ŽEMEP-WMO, 1997; Eliassen and Saltbones, 1983; Tarrason and Tsyro, 1998; Pacyna et al., 1991. and several national and international efforts ŽBerdowski et al., 1998; EPA, 1996a,b; EU, 1996, 1997; Quality of Urban Air Review Group, 1996; EU Position Paper on Particles, 1998.. The main objectives of the EMEP programme are to provide quantitative informa-

0048-9697r02r$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 0 1 . 0 0 9 3 2 - 9

210

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

tion on the transport of air pollutants across national boundaries and on the deposition and concentration levels caused by this transport. EMEP’s work in the past has been related to acid precipitation and photochemical oxidant formation, and lately heavy metals and persistent organic compounds have been included ŽEMEPWMO, 1997; Eliassen and Saltbones, 1983; Tarrason and Tsyro, 1998; EMEP, 2000a,b.. A total of over 150 stations across 35 countries covering the whole of the European continent have produced data for monitoring the concentrations and trends in acidifying and eutrophying atmospheric pollution. These data present a very valuable database for the characterisation of Europe’s atmosphere and the potential changes. Long term time series from the EMEP monitoring programme have been employed closely in the development of several international protocols, on reduction of acidifying and eutrophying emissions. The EMEP aims also to consider progress towards targets of environmental quality, and observations from the EMEP monitoring programme over the past two decades provides information from which the degree of such improvement can be estimated. Assessment is thus made as to whether the underlying intentions of the agreed emission reductions from the European countries have been met. Emissions of pollutants rise up in the air due to buoyancy effects, advect downwind, and disperse horizontally and vertically due to turbulence field and prevailing meteorological patterns. In recent years there has been an extensive research focus on particulate matter ŽPM. ŽMcMurry, 2000; Turpin et al., 2000; Heintzenberg et al., 1998; Andreae and Crutzen, 1997. mainly because of serious public health risks for susceptible members of the population and risks to sensitive ecosystems. The relationship between exposure to airborne PM and resulting health effects is the subject of an ever-increasing number of studies ŽPope et al., 1995; Schwartz and Dockery, 1992; Schlesinger, 1995.. While epidemiologic studies suggest an association between ambient PM concentrations and increased human morbidity and mortality ŽSchwartz et al., 1996; Wilson and Suh, 1997., new toxicological studies start to present

potential biological explanations for this observed association ŽEPA, 1996a.. Another implication of ambient PM concentrations is connected with the wet and dry deposition of sulfate and nitrate particles to the earth’s surface leading to acidification and eutrophication. Eutrophication is becoming a serious threat to coastal environments ŽEMEP-WMO, 1997; Pelley, 1998.. In addition, visibility degradation is one of the most readily perceived impact of particles since they absorb and scatter the light and therefore reducing visibility ŽTrijonis et al., 1991; Hegg et al., 1993.. Particulate matter also influences the climate directly Žthrough scattering and absorption of the solar radiation. and indirectly through the formation of cloud condensation nuclei ŽCharlson et al., 1992; Hoffmann et al., 1997.. A significant economic effect connected with PM is the soiling of man-made surfaces. The process of cleaning, painting and repairing exposed surfaces becomes an economic burden. Acid particles can severely deteriorate art works and historic monuments ŽBaedecker et al., 1992; Ligocki et al., 1993.. The monitoring of PM in the EMEP framework is mainly concerned with health effects on humans and eutrophication᎐acidification Žagriculture and forestry impacts. of the European area. A comprehensive and realistic monitoring strategy of the EMEP programme was proposed during an EMEP-WMO workshop ŽEMEP-WMO, 1999.. Topics such as the geographical coverage of the monitoring stations, the type of measurements, temporal coverage and recommended sampling and analysis methodology were considered. In particular it is important to note that many measurement sites in Europe are already both EMEP and GAW ŽGlobal Atmospheric Watch. programmes of the World Meteorological Organisation ŽWMO.. Atmospheric aerosol mass size distribution has generally a minimum in the range between 1 and 3 ␮m. It is common also to define the particle size into three modes, the nuclei mode Žparticles with diameter below 0.1 ␮m., accumulation mode Žwith diameter between 0.1 and 1 ␮m. and the coarse mode Žwith diameter greater than 1 ␮m up to 100 ␮m.. The transport of PM in the atmosphere is

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

similar to that of gaseous pollutants for the fine particle fraction but deviates at larger sizes Žcoarse particles. due to deposition processes. The longrange transport of PM contributes significantly to the background particle mass and number size distribution. However, there is still considerable debate among the scientific community regarding the vertical distribution and the spatial and temporal scales of atmospheric particle transport. Airborne particulate matter is a complex mixture of many different chemical species originating from a variety of sources. Composition, morphology, physical and thermodynamic properties of PM varies in different geographical places and has a seasonal variability ŽFinlayson-Pitts and Pitts, 1986; Seinfeld and Pandis, 1998; EPA, 1996a,b; EU Position Paper on Particles, 1998; Alpert and Hopke, 1981; Hoffmann et al., 1997.. An important characteristic of atmospheric particulate matter is the large variation in size ranging from tens of micrometers to a nanometre size ŽEPA, 1996a; Covert et al., 1992; Clarke, 1992.. For example combustion-generated particles Žvehicle emissions, power generation. range in size between 0.003 and 1 ␮m. Pollen and soil dust is composed of particles mainly above 2 ␮m, whereas fly ash from coal combustion produces particles ranging from 0.1 to 50 ␮m ŽEPA, 1996a; Seinfeld and Pandis, 1998.. In addition, aerosols in the atmosphere undergo changes in their chemical composition and size such as nucleation Žnew particle formation., condensation, evaporation, coagulation, deposition Žboth wet and dry., activation due to water and other gaseous species and aqueous phase reactions ŽSeinfeld and Pandis, 1998; Finlayson-Pitts and Pitts, 1986.. The lifetime of particulate matter in the atmosphere is between a few days Žcoarse particles. to a few weeks Žfine particles.. The relative long residence times of fine particles compared to coarse ones results in small differences between the average total mass of PM 2.5 between urban and non-urban continental aerosols ŽHeintzenberg, 1989.. In the central and northern part of Europe anthropogenic sources are dominant because of urbanisation and the large number of vehicles and stationary combustion sources Žindustrial and residential . ŽEU Position Paper on Particles,

211

1998.. Natural sources of primary aerosols in Europe include sea spray, fugitive dust Že.g. soil resuspension by the wind., long-range transport of Sahara dust, volcanic and biogenic emissions. For the southern European Countries Saharan dust and forest fires are important aerosol sources ŽRodriguez et al., 2001.. In arid areas of southern Europe resuspended dust has an important contribution to aerosol mass ŽEU Position Paper on Particles, 1998.. Furthermore, at coastal regions the sea spray contribution is important, especially during high wind velocities ŽEU Position Paper on Particles, 1998.. There is a consistent pattern of geographical variability in Europe with lower regional background concentrations of PM 10 in the far north and higher concentrations in the southern countries ŽEU Position Paper on Particles, 1998.. This is partly due to natural emissions of unsaturated hydrocarbons Žincluding isoprene., which are highly reactive. However, in some parts of Europe the majority of monitoring data is from urban networks and there is no systematic monitoring program with representative rural sites in most countries. In addition, many research studies have been performed in north-western Europe, where aerosol concentrations between urban and nonurban areas on an annual basis do not differ by more than 20% Že.g. Van der Zee et al., 1998.. This can be attributed to widespread emissions of precursors of PM from a dense population area and a characteristic transport distance for PM which is larger than the distance between the measurement sites highlighting the importance of long-range transport of air pollutants. Furthermore, in a study on wintertime concentrations comparing PM 10 and black smoke in 14 urban and 14 non-urban locations in Europe, indicate a relative small difference Žon average 22% for PM 10 and 43% for black smoke. ŽHoek et al., 1997.. Harmonised measurements for determining particle size are essential for obtaining comparable data ŽCEN, 1998.. Similar observations on the regional character of particulate matter are also reported in the United States ŽEPA, 1996a,b.. It is evident that a more detailed chemical and physical characterisation of ambient aerosols beyond PM 10 mass measurements is needed.

212

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

This framework has been already discussed and it is planned to establish a number of ‘superstations’ in Europe together with other scientific organisations and programmes ŽEMEP-WMO, 1999.. The scope is the determination of specific organic compounds, size distribution measurements, detailed size fractionated chemical speciation, optical properties, water uptake, cloud condensation nuclei ŽEMEP-WMO, 1999.. Here we concentrate mainly on the currently available measurement data of particulate matter in Europe inside the EMEP monitoring network, recommended measurement methods, as well as at providing recommendations for future work.

2. Particulate matter measurement techniques in the EMEP framework A number of different methods have been used to quantify the concentration of particulate matter in ambient air. These may either be based on the sampling of air through a suitable filter and subsequent weighing, or on other methods, such as by monitoring the absorption of ␤-radiation through the filter ŽWedding and Weigand, 1993., or the monitoring of frequency of an oscillating tapered element with a filter at the tip ŽTEOM. ŽPatashnik and Rupprecht, 1991.. Other methods are based on light scattering, or counting of particles with different mass through electrical charging and determining their mobilities. All these methods can and will give different results, and the European Standardisation Organisation ŽCEN. has therefore designed three different reference methods for determination of the mass of particulate matter with aerodynamic equivalent diameter - 10 ␮m ŽCEN, 1998.. Instrumentation for aerosol measurements can be classified based on their capacity to resolve size, time and composition ŽMcMurry, 2000.. Continuous measurements of integral properties Že.g. number, mass, surface size distribution. include instruments such as Condensation Nuclei Counters ŽCNC., epiphaniometer, integrated nephelometer and photoacoustic spectrometer. Size distribution measurements at short time interval Žseveral minutes. include instruments such

as electrical mobility analysers, diffusion battery and aerodynamic particle sizers. Air intake efficiencies, loss of particles in tubes and connections and calibration accuracy are points that have to be checked for reliable results. More recent realtime measurement methods include the individual particle composition with mass spectrometers. Methods that are based on ␤-radiation through the filter or the TEOM are also used extensively for continuous monitoring of aerosol mass. Evaporation losses due to heating and underestimation of mass on filters are problems to be considered. Measurements of the aerosol size distribution with time-integrated measurements are mostly performed with the use of impactors. Problems in sampling efficiency of impactors arise due to particle bounce, overloading and interstage losses ŽEPA, 1996a.. Gravimetric measurements with specific cut-off sizes are also widely used in monitoring studies. Loss and absorption of gases from and to filters is also a problem that has to be considered in these measurements. Methods of particulate mass measurements and results from comparative testing of several methods Žgravimetric, ␤-radiation through the filter, and the TEOM for PM 10 and PM 2.5 . have been reviewed in a workshop meeting by the WHO Ž1999.. Direct recording instruments can be used if they have been shown to provide consistent results compared with gravimetric methods. The main problem with the instruments that are monitoring the frequency of an oscillating tapered element is that heating to remove water contained in the aerosol particles will also lead to the evaporation of ammonium nitrate and volatile organic compounds. When these substances constitute a large fraction of the aerosol mass, some of the monitoring instruments will give erroneously low readings. The gravimetric method will also be subject to sampling artefacts, and losses of ammonium nitrate may also occur during sampling and when filters are suspended in a weighing room for 48 h at 25⬚C and 50% RH. These artefacts will be addressed in separate tests and investigations that are now going on in several European countries in preparation for a new European PM 2.5 standard. According to Community Directive 96r62rEC

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

on ambient air quality assessment and management and Directive 1999r30rEC relating to the establishment of limit values, inter alia also for particulate matter, all member states of the EU are obliged to report PM 10 and PM 2.5 data. In order to assess air quality across Europe on a consistent basis the measurements need to be performed with standardised techniques. Therefore the European Commission has mandated the European Standardisation Committee ŽCEN. to establish reference methods for measurement of PM 10 and PM 2.5 and to define requirements for alternative methods to be considered as equivalent methods. The work for PM 10 is now completed and has led to the standard EN-12341 while the work on PM 2.5 is still in its beginning phase. As for PM 10 , a manual gravimetric reference method will most probably be defined for PM 2.5 . With this procedure Europe is in line with the corresponding methods defined by the EPA for the USA. The measurements of PM 2.5 are very important since they are highly correlated with health effects ŽWilson and Suh, 1997.. Ozone measurements at the same sites where particulate matter is monitored is highly desirable because of the effects of ozone to human health and ecosystems as well as due to correlation between ozone and fine particles during photochemical episodes. The EMEP programme has several sites where ozone is monitored ŽEMEP, 2000b.. Furthermore, at the PM 10 sites there is also monitoring of ozone. Following the recommendations made in the EMEP-WMO fine particulate workshop ŽEMEPWMO, 1999. EMEP should give first priority to PM 10 measurements. For this purpose the gravimetric method is the preferred method, particularly because the filters allow subsequent chemical analysis for quantification of different compounds. Monitors are acceptable if they have been shown to give equivalent results for the specific site and for all seasons. EMEP sites should also determine secondary inorganic particulate matter, i.e. ammonium sulfate and ammonium nitrate, as well as other water-soluble ions when these make up a significant part of PM 10 mass. Limiting the sampling to particles smaller than 10 ␮m is justified by the wish to only study

213

particles that have an effect on human health. This has the added advantage that it reduces the effects caused by resuspension of soil particles from the ground, and it also removes artefacts in sampling at high wind speeds. However, comparison of measurements of PM 10 and total suspended particulate in central and southern Europe show that the contribution of particles larger than 10 ␮m to TSP is moderate in most places in Europe ŽEMEP-WMO, 1999.. Quartz fibre filters are specified for the gravimetric determination of particulate matter in the EMEP sites for PM 10 measurements. These have good retention properties with low pressure drop even for very high sampling rates. Absorption of gases on the filter material is also a limited problem. Water vapour retention by the quartz fibre material requires very careful conditioning of the filters before weighing. CEN Ž1998. specifies at least 48 h at 25 " 2⬚C and 50 " 5% relative humidity. This should also remove any water of hydration associated with water-soluble salts such as ammonium nitrate and ammonium sulfate. However, the conditioning may also result in loss of volatile components, such as ammonium nitrate and absorption of organic vapours on the quartz surface. Additional recommendations identify that measurements of particles with an aerodynamic diameter less than 2.5 or 1 ␮m should be carried out in the near future when the definition of the European reference method is in place. More detailed size fractionated chemical speciation would be desirable and should be done in the context of scientific projects. Furthermore, for chemical characterisation, determination of elemental and organic carbon is highly desirable. The subsequent determination of organic and elemental carbon by thermo-desorption and oxidation is subject to artefacts, and care has to be taken to avoid results that are not comparable. The EMEP programme currently uses a method for ECrOC analysis that is based on the equipment described by Birch and Cary Ž1996.. Elementary carbon is determined by successive volatilisation and oxidation of the sample and the determination of the evolved CO 2 is performed either directly or after conversion to CH 4 by a

214

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

flame ionisation detector ŽFID.. This procedure also gives the total carbon content, and a quantification of the amount of organic materials through the organic carbon content of the aerosol particles. This method is not free of artefacts, particularly the charring or incomplete removal of organic compounds may lead to the overestimation of EC. To compensate for this, optical detection of a darkening of the filter during the last stage of the OC volatilisation is performed ŽChow et al., 1993.. Another aim in the EMEP programme is the establishment of ‘Superstations’ together with other scientific organisations and programmes. These could be used for a number of chemical and physical measurements that go beyond the scope of the ‘normal’ EMEP site, e.g. for determination of some of the organic compounds, size distribution, detailed size fractionated chemical speciation, optical properties, water uptake, cloud condensation nuclei, vertical distribution and better time resolution Ž1 h. for certain parameters. The EMEP measurement strategy for future monitoring of ambient particulate matter is determined in three levels Žlevel I, II, III. ŽEMEP, 2001.. At level I there is an urgent need to perform PM 10 measurements at least at one EMEP station in each country following gravimetric methods ŽCEN, 1998.. The ratio of elementary to organic carbon ŽECrOC. is also recommended to be determined at least once every week from each station in the start phase. At the same time measurements of soluble base cations, sodium, potassium, calcium and magnesium as well as chloride should be initiated. In the start phase the measurements have to be carried out together with the ECrOC analysis. For the EMEP stations that determine sulfate, nitrate and ammonium the base cations and chloride should be determined on a daily basis. In the level II denuder measurements are required to distinguish between the concentration of gaseous and particle-bound nitrates and ammoniarammonium compounds. There is urgent need to map the spatial and temporal distribution of the nitrate aerosol over Europe due to the importance of nitrates to radiative forcing, acidification and eutrophication. Furthermore, nitrate

aerosols are an oxidation product of NO x and are highly correlated with ozone. Measurements of mineral dust with a 24-h sampling period are also foreseen ŽEMEP, 2001.. Finally, in level III EMEP is foreseen to co-operate with research projects across Europe to obtain additional data on aerosol size distribution with detailed chemical speciation for each size range and detailed organic carbon speciation. A Žspatial. speciation of the organic fraction is of great interest, e.g. as emission indicators and for better understanding the effects of particle size chemical characteristics on human health.

3. Particulate measurements inside the EMEP framework Switzerland, Germany, Italy and Spain only began reporting measured concentrations of suspended particulate matter at their EMEP stations from 1997, using gravimetric measurement methods. Data are available for the period 1997᎐1999 and data for the year 2000 will soon be reported. Germany reports total suspended particle ŽTSP. data, using a special high-volume sampler ŽLIS.. These measurements were started before 1980 at several Umweltbundesamt stations. During 1999 the samplers were modified with new PM 10 inlets to meet the EN12341 standards for PM 10 measurements. Switzerland reported PM 10 data since 1997, before 1997, TSP data have been reported. Extensive co-located measurement data with TSP, PM 10 and PM 2.5 samplers are available for the Swiss stations ŽEMEP-WMO, 1999.. Spain has also reported TSP data since 1989. Italy reports TSP data only from one station ŽIspra. and from the year 2000 PM 10 and PM 2.5 data are also available. The Netherlands have measured particulate aerosol mass at several sites, including the EMEP sites Vredepeel ŽNL09., Eibergen ŽNL02. and Witteveen ŽNL07.. A monitoring instrument ŽFAG FH 621-N. based on ␤-ray absorption is used, this gives systematically too low results according to work performed in Germany ŽEMEPWMO, 1999.. Measurements that are not following the EN-12341 measurement standard for

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

PM 10 can be accepted if it is documented that results equivalent to the standard are produced or if constant correction factors are applied. In order to compensate for the differences with the ␤-ray absorption method the results have been multiplied by a constant factor of 1.33 ŽBuijsman and van Elzakker, 1996.. The same instrumentation is used at the German UBA stations, without corrections the results are typically 30᎐40% lower than the results obtained with the LIS sampler. Measurements with a ␤-ray absorption instrument are also carried out in the Czech Republic at several rural stations, including Kosetice ŽCZ03.. No correction factors are being used in the Czech network. Some rural concentration data are also available from the United Kingdom, these measurements have been made with TEOM instruments, which may underestimate PM 10 concentration relative to the reference methods by as much as 30%. This is due to volatilisation of the aerosol mass sample in the filter. Again, German investigations support this ŽEMEP-WMO, 1999., but there is a reasonably good correlation between TEOM and PM 10 on a daily basis, suggesting that a constant correction factor may be acceptable. Annual averages of PM are presented in Table 1. Details on the station characteristics, representativeness and operational status can be found in the EMEP web site Žwww.emep.int.. The measurements within EMEP are carried out by national laboratories, reporting the results to a common data base at the Chemical Coordination Centre ŽCCC. which is located at the Norwegian Institute for Air Research ŽNILU.. Quality assurance is carried out both at the national level and by the CCC to ensure satisfactory data quality. This applies both to individual samples and particularly to long-term aggregated values, such as seasonal or yearly mean values and trends ŽEMEPrCCC 1r95, 1996.. It is particularly important to avoid errors which may result in systematically too low or too high results, and undefined changes in the data quality over time, which may cause problems in trend analyses. The representivity of a site is a highly relevant question for a measurement network such as EMEP. This can only be determined in relation to the purpose of the measurements. For the

215

Table 1 Measured aerosol particle concentrations at some EMEP sites Žavailable data from 1997. Site

SPM Ž␮grm3 .

Method

CH01 Jungfraujoch CH02 Payerne CH03 Tanikon ¨ CH04 Chaumont CH05 Rigi

3.5 26.3 27.3 13.9 14.4

PM10

DE01 Westerland DE02 Waldhof DE03 Schauinsland DE04 Deuselbach DE05 Brotjacklriegel DE07 Neuglobsow DE08 Schmucke ¨ DE09 Zingst

26.2 24.3 13.3 20.7 15.4 23.1 17.3 23.8

LIS

ES01 Toledo ES03 Roquetas ES04 Logrono ES05 Noya ES06 Mahon ES07 Viznar IT04 Ispra

19.7 44.4 33.0 24.5 28.4 30.9 46.8

TSP

NL02 Witteven NL10 Vreedepeel SE12 Aspvreten

Ž35. Ž44. Ž9.5.

␤-abs. ␤-abs. TEOM

SPM, suspended particulate matter; CH, Switzerland; DE, Germany; ES, Spain; IT, Italy; NL, The Netherlands; SE, Sweden.

EMEP the site must be positioned in such a way that the air quality and the precipitation is representative of a larger region. The size of this area is determined by the variability of the air and precipitation quality, and the desired spatial resolution in the concentration and deposition fields. The purpose of EMEP is to provide the European countries with information on the deposition and concentration of air pollutants, as well as on the quantity and significance of the long-range transmission of pollutants and fluxes across national boundaries. The size of the site’s area of representativeness should be larger than the size resolution of the atmospheric dispersion models which are available for the evaluation of the long-range transmission and deposition of air pol-

216

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

lutants. EMEP models and emission surveys have up to now employed grid sizes of 150 = 150 km2 , this spatial resolution is now being improved to 50 = 50 km2 . When the major part of the emissions influencing the air quality in the area are situated outside that area, selection of the site mainly involves consideration of the effects of the immediate surroundings and emissions within the nearest 20 km. These local emissions should not be allowed to result in unrepresentative measured air concentrations or precipitation chemistry at the site, which means that their influence must be evaluated and compared with the measurements. The site must be representative also with respect to exposure to air mass. Details on the measurement strategy in EMEP can be found in EMEPrCCC 1r95 Ž1996.. Details of the position and characteristics of the stations can be also found in EMEP reports ŽAas et al., 1999.. Measurements of suspended particulate matter ŽSPM. from the sites Aspvreten ŽSE12., Kollumerwaard ŽNL09. and Vredepeel ŽNL10. have been obtained directly from national authorities. The measured values may not be directly comparable, because of different measurement methods and procedures. This is partly because of differences in the air intakes, but also because of other sampling artefacts, particularly evaporation of ammonium nitrate and other volatile constituents from filters during sampling. This is particularly serious in the heated filters used in TEOM and ␤-absorption instruments ŽAllen et al., 1997; Patashnik, 1998.. In Figs. 1᎐3 time series of measured PM concentrations for the 1997 and 1998 from a number of EMEP stations in Switzerland, Germany and Spain are presented. Concentrations in Switzerland show an annual average of 21 ␮grm3 Žexcluding the Jungfraujoch measurements, height of 3500 m. similar to Germany Ž24 ␮grm3 .. Spain shows higher concentrations with an annual average of 30 ␮grm3 in accordance with expectations due to the dry climatic conditions. Some Spanish sites have high summer PM concentrations, but there is no consistent seasonal variability. For the Jungfraujoch the concentrations are generally very low Ž␮grm 3 ., but high concentrations are

Fig. 1. Time series for ambient particulate matter concentrations Ž␮grm3 . in the EMEP stations of Jungfraujoch and Payerne ŽSwitzerland. during 1997᎐1998.

observed in cases when the atmospheric boundary layer reaches the elevation of the station. In Switzerland the PM concentrations are generally higher in winter than in the other seasons, due to inversions during winter and better mixing during the warmer seasons. A spatial distribution of PM 10 ŽGermany, Switzerland. and TSP ŽSpain, Italy. annual average measurements from the EMEP monitoring programme for 1998 is presented in Fig. 4. Lower concentrations are observed in Switzerland and especially at the Jungfraujoch station which is usually located above the atmospheric boundary layer. The TSP levels in Spain Žfour stations. and Italy Žone station. show annual average values

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

217

the data using the procedure recommended by CEN and the CEC ŽCEN, 1998.. Based on experience we can therefore expect that there could be systematic differences of the order of "20᎐30% between the measurement series in different countries ŽWHO, 1998.. These results show, however, the order of magnitude of the aerosol particle mass concentrations, and that the difference between rural and urban concentration levels is rather small. This is particularly the case in continental Europe, where the regional emission levels and the secondary aerosol production appear to be more important than the urban emissions. Total particulate sulfate and nitrate are the dominant components, with ammonium nitrate prevailing in Western Europe and particulate sul-

Fig. 2. Time series for ambient particulate matter concentrations Ž␮grm3 . in the EMEP stations of Westerland and Langenbrugge ŽGermany. during 1997᎐1998.

close to 30 ␮grm3. Detailed temporal variation of the aerosol concentrations in the EMEP stations for 1997᎐1998 can be found in the report by Lazaridis et al. Ž1999.. Particulate matter annual average concentrations in the EMEP stations in 1999 are presented in Fig. 5. The concentrations are very similar to 1998 levels. Spain included from 1999 additional number of stations from its national monitoring programme to the EMEP. For comparison with the EMEP data for 1997 we present in Table 2 data on ambient suspended particulate mass from the European Environmental Agency’s AIRNET database ŽLarssen and Lazaridis, 1998. and from the EMEP sites. Only some of these sites use the recommended CEN12341 reference method, or have corrected

Fig. 3. Time series for ambient particulate matter concentrations Ž␮grm3 . in the EMEP stations of Toledo and Roquetas ŽSpain. during 1997᎐1998. Maximum value for Toledo in 1998 was 342 ␮grm3.

218

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

fate Žmainly as ammonium sulfate . prevailing in Eastern Europe Žsee Van der Zee et al., 1998; Zappoli et al., 1999.. Together with the secondary inorganics, primary aerosols, largely organic and elemental carbon, minerals and crustal material, explain a major part of the particulate mass concentration levels in Table 2. Recent studies Že.g. Heintzenberg et al., 1998. emphasise the importance of organic compounds, whereas minerals from fly ash apparently have decreased strongly. With respect to primary particles, the emission estimates at present are relatively uncertain, and the measurements do not provide any validation of the emission data, nor of the model estimates. However, it is difficult to derive significant conclusions on the actual variations in particulate matter observed over Europe because differences in the sampling and analytical methods can be of the same order of magnitude as regional variations. Here we include PM air quality data for 1997

in Europe transmitted by countries on a voluntary basis in the framework of the ‘Exchange of Information’ Decision Ž97r101rEC. as well as of data from EMEP sites. This kind of data reporting is essential for summarising the collected information and outlining the underlying trends in air quality in Europe ŽLarssen and Lazaridis, 1998.. The European Topic Centre on Air Quality ŽETC-AQ., under contract from the EEA ŽEuropean Environment Agency. is managing the database system AIRBASE. In this chapter we present summarised air quality data on PM concentrations from the AIRBASE and EMEP databases. In Fig. 6 we illustrate the measured levels of PM 10 at rural, urban and roadside sites. Data from rural areas are from both the AIRBASE and EMEP databases. Different measurement methods have been used to quantify the particulate mass concentration levels in this figure. The data in the AIRBASE database are obtained from national monitoring programmes

Fig. 4. Annual averages of the PM 10 and TSP concentrations Ž␮grm3 . from the EMEP monitoring framework for 1998.

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

in European countries. The sites from which data are selected have been checked for both quality assurance and representativeness. A detailed description on the criteria followed for site selection is presented by Larssen and Sluyter Ž1999.. The term ‘high’ refers to the highest value among the stations and the number at the top of each bar indicates the number of measuring stations in each country. In the Netherlands and the Czech Republic beta-gauges are used, while the UK uses TEOM instruments. Both beta-ray absorption monitors and TEOM give results that are too low compared to the CEN 12341 gravimetric reference method for PM 10 , The Netherlands uses a constant factor to correct for this instead. The figure, however, clearly indicates that the rural concentration levels of particulate mass are not much lower Ž10᎐30%. than the urban and roadside concentrations.

4. Aerosol characterisation of PM inside the EMEP framework As explained in the previous section few EMEP sites report TSP or PM 10 mass concentrations using the gravimetric method. However, chemical analysis for a number of particulate matter components is carried out in the EMEP network. Sulfate is determined at the majority of the sites, ammonium and nitrate are at present often reported as the sum of ammonia and ammonium, and the sum of nitrate and nitric acid, respectively. However, experience has shown that ammonium associated with particles is generally explained as ammonium sulfate and ammonium nitrate ŽSemb et al., 1998.. This may not be entirely justified, particularly not in Spain, where a significant fraction of both sulfate and nitrate may also be associated with alkaline soil dust particles. The ratio of nitrate associated with particles to the sum of nitrate and Žgaseous. nitric acid has been studied in measurement campaigns carried out within EMEP in 1993 ŽSemb et al., 1998; see also Fig. 7.. The measurements carried out during four periods of minimum 2 weeks, in December 1992, and March, June and September 1993. The results from the measurement campaign

219

also show that the fraction of gaseous nitric acid is typically 20᎐30%, relative to the sum of nitrate and nitric acid. The lowest relative concentration of nitric acid is found in areas with high groundlevel concentrations of ammonia, and relatively low temperatures. It was found that ammonia was not generally limiting the formation of ammonium nitrate, since the sulfur dioxide emissions and sulfate concentrations have been reduced. Ten sites took part in the pilot measurements as follows: DK33 Lille Valby ŽDanmark., CH2 Payerne ŽSwitzerland., FI9 Uto ŽFinland., IT4 Ispra ŽItaly., HU2 K-Puszta ŽHungary., NO1 Birkenes ŽNorway., RU14 Pushkinskie Gory ŽRussian Federation., SE2 Rorvik, SE12 Aspvreten ŽSweden.. There is relatively good correlation between observed TSP concentrations and concentrations of aerosol sulfate Žand nitrate . at many of the above sites, even if the sulfate and nitrate concentrations account only for 20᎐40% of the aerosol mass. This indicates that the aerosol particle mass concentrations are part of a long-range transported or regional air pollution components, which also includes particles which are emitted from various anthropogenic sources, such as stationary combustion sources, mobile sources and industrial processes. Closing the mass balance also requires the determination of elementary carbon Žsoot particles. and organic compounds. The latter consist of a very large number of individual compounds, and the identified compounds or groups of compounds only account for some 15᎐20% of the organic fraction Že.g. Rogge et al., 1993; Liousse et al., 1993.. Elementary carbon is typically 5᎐10% of total carbon. Examples of chemical composition of fine particles at three rural stations, in Italy, Hungary and Sweden are given by Zappoli et al. Ž1999.. Two of these sites ŽK-puszta and Aspvreten. are also EMEP sites. Care should be taken with respect to the representativity of these results, since they only represent 2 months of sampling at three sites. The contribution of both water-soluble and -insoluble organic carbon to the fine particulate mass is large. A very high contribution of water-soluble organic carbon to

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

220

Fig. 5. Annual averages of the PM 10 and TSP concentrations Ž␮grm3 . from the EMEP monitoring framework for 1999.

the mass at Aspvreten may be related to the low fine particulate mass concentration. The importance of organic compounds for the particulate mass is also emphasised by Heintzenberg et al. Ž1998., who reported on extensive measurements from Melpitz, near Leipzig in south-eastern Germany. The concentration of coarse particles consisting mainly of inorganic minerals and fly ash particles has decreased strongly since 1993, and now represent only approximately 20% of total particulate mass, or PM 10 on an annual basis.

Only limited information is available with respect to the other chemical components in the aerosol particles. The contents of inorganic elements have been studied with various instrumental methods of chemical analysis, mainly focussed on the chemical signatures and the occurrence of particular trace elements which may be used to indicate source types and geographical regions. The elements which contribute most to the particulate mass are the elements most common in the Earth’s crust, namely silicon, aluminium, iron, calcium, a.o. Given that the concentration of these

Table 2 Some annual averages of measured particulate mass concentration levels Ž1997.

Country PM10 , ␮grm3 No. sites

EMEP sites

AIRBASE data

Rural

Rural

CH 21 4

DE 24 5

ES 30 4

CZ 25 19

Urban NL 39 9

GB 14 3

CZ 38 34

BE 32 6

NL 41 6

GB 23 36

IE 17 1

NO 19 3

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

221

Fig. 6. Annual averages of the PM 10 concentrations in different countries in Europe for various site types.

three or four elements have been determined, a good estimate of the contribution from inorganic primary particulate emissions may be made. Seasalt concentrations are similarly inferred from the measured concentrations of sodium and chloride.

Chloride is often depleted in aerosol filter samples because of interactions with acid aerosols, or nitric acid. Fig. 8 shows the development with time of the concentrations of some elements at Birkenes,

Fig. 7. Concentrations of particulate sulfate, nitrate and ammonium from denuder measurements during the EMEP measurement campaign. Measurements were performed for periods of 4᎐5 weeks. Concentrations are given in weight equivalents for comparison with PM 10 concentration levels Ž␮grm3 ..

222

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

Norway from 1973 to 1996 based on EMEP data. It is seen that the concentrations have been decreasing strongly, particularly for sulfate and for the inorganic mineral elements. This is in accordance with emission reductions, and with similar observations concerning the decrease in deposition of base cations in precipitation ŽPeirson et al., 1973; Hedin et al., 1994.. The total mass of the inorganic mineral elements has been estimated, together with the mass of ammonium sulfate estimated from the sulfur content, and the mass of sea-spray aerosol particles. When these are compared with the total mass, determined by weighing of the membrane filters prior to the chemical analyses, a significant part remains, typically approximately 30᎐40% of the total mass, which consists of organic material and elementary carbon. Optical absorption measurements indicate that the elementary, or black carbon concentration level is approximately 0.5᎐1 ␮grm3. This corresponds well with other measurements of the chemical composition of aerosol particles in Europe ŽHeintzenberg et al., 1998.. Of particular interest is the relatively high pro-

portion of water-soluble organic material, which may be attributed to chemical transformations in the atmosphere, either in the particles or by transformation of gaseous organic material that is subsequently absorbed in the aerosol particles. The formation of secondary organic aerosol has been the subject of extensive laboratory studies, and occurs both with the oxidation of naturally emitted terpenes and with anthropogenic VOC, particularly VOCs with high molecular weight. Of the groups of compounds which can be identified, hydroxylic, carbonylic and carboxylic acid derivatives of both aliphatic and aromatic compounds are all present. The content of organic material in aerosol particles and their relationship with both primary emissions and atmospheric processes is a considerable challenge, more so because the other components of atmospheric particulate matter are declining because of emission reductions. Heintzenberg et al. Ž1998. have summarised the experience from research and observations of atmospheric particles in the Leipzig area, and find that the fine aerosols Ž- 1᎐2 ␮m. consist almost

Fig. 8. Chemical speciation Žngrm3 . of suspended particulate matter in Birkenes ŽNorway. in the period 1973᎐1993.

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

223

Fig. 9. Measurement results from Lille Valby ŽDK33..

exclusively of organic and secondary inorganic particulate material. Previous studies indicate ŽSemb et al., 1995, 1998. that the contribution from secondary inorganic aerosol components, namely ammonium sulfate and ammonium nitrate, is between 20 and 40%, and that 30% is rather typical of northern Europe. If the sampling was restricted to fine particles, i.e. particles less than 2.5-␮m-diameter, the weight percentage of the secondary inorganic aerosol component should increase.

Methods used to obtain aerosol mass concentrations by sampling on filters will always have some limitation. Loss of volatile species during sampling has been mentioned, another artefact is related to water uptake by deliquescent salts. Most water-soluble inorganic salts will take up water at relative humidity above 60᎐70%, and aerosols will usually contain substantial amounts of water, which will be retained even at lower relative humidity, because crystallisation of the pure salts is inhibited. The preferable method for

224

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

ŽSE02.. Fig. 10. Measurement results from Rorvik ¨

measurement of PM 10 mass in Europe within EMEP is determined according to EN 12341 ŽCEN, 1998.. The sampling period should be 24 h and the samples should be changed daily together with the other sampling devices between 07:00 and 09:00 h local time ŽEMEP, 2001.. This measurement should give the PM 10 mass and should be in accordance with the EN 12341. A full chemical analysis of particles sampled on filters is also not straightforward, mainly because of the many different organic compounds. The

quantification of elementary carbon and organic compounds ŽECrOC. in aerosol particles is of considerable interest. The ratio between EC and OC is often used as a valuable tool for the elucidation of the origin of the air masses investigated ŽLiousse et al., 1993.. Determination of carbon following stage-wise thermo-desorption and oxidation may be used to quantify the amounts of elementary and organic carbon Že.g. Cachier et al., 1989.. It is more technically demanding to separate different classes of com-

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

pounds by selective extraction. The fraction of the organic compounds which are water-soluble may be quantified in this way, however. Good analytical methods exist for the determination of inorganic elements, and there is a large body of information concerning the inorganic elemental composition of aerosol particles, dating from the 1970s. Že.g. Pacyna et al., 1984a; Lannefors et al., 1983; Amundsen et al., 1992; Pakkanen et al., 1996.. These studies also show, however, that the concentration of inorganic minerals in

225

aerosols have declined over this period mainly because of emission reductions from process industries and solid fuel combustion. 4.1. Sulfate, nitrate and ammonium measurements Representative results from individual measurements concerning average concentrations of gaseous nitric acid, particulate nitrate, gaseous ammonia, and particulate ammonium, on a daily basis are shown in Figs. 9᎐13. The concentrations

Fig. 11. Measurement results from Aspvreten ŽSE12..

226

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

of sulfate aerosol are also given when available. A detailed study on nitrogen pilot measurements in the EMEP programme is presented by Semb et al. Ž1998.. The measured concentrations of gaseous nitric acid were generally much lower than the concentrations of particulate nitrate. At the sites IT04 ŽIspra., CH02 ŽPayerne. and DK33 ŽLille Valby., relatively high concentrations of gaseous ammonia are effective in converting nitric acid to ammonium nitrate and suppressing ammonium ni-

trate dissociation. The relative amount of gaseous nitric acid to nitrate in particles increases during the summer months at these sites. Relatively high concentrations of gaseous ammonia are also measured at HU02 ŽK-puszta., but here the measured concentrations of nitric acid are comparable to the concentration of nitrate particles. Concentrations of ammonia are low, but still measurable at SE02 ŽRorvik ¨ . and NO01 ŽBirkenes. as shown in Figs. 10 and 12. However, the concentrations of gaseous nitric acid relative

Fig. 12. Measurement results from Birkenes ŽNO01..

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

to the concentrations of particulate nitrate at these sites were on average only 21 and 16%. At SE12 ŽAspvreten. Žsee Fig. 11. concentrations of ammonia were very low. Nitric acid and nitrate concentrations were also low, however. Particulate ammonium is usually associated with particulate sulfate as ammonium sulfate. Ammonium nitrate and gaseous ammonia can only occur when sulfate particles are fully neutralised Žas ŽNH 4 . 2 SO4 .. Comparison of the concentrations of particulate sulfate, particulate ammonium and particulate nitrate can therefore be

227

used to test if the particulate nitrate is in the form of ammonium nitrate or as other salts. It is seen that there is generally good correlation between particulate ammonium and particulate sulfate. At IT04 ŽIspra. the measurements indicate that ammonium nitrate is an important ammonium compound, but the excess of ammonium over the amount corresponding to ammonium sulfate is often too low to explain the particulate nitrate concentrations. As there is also an excess of ammonia, chemical absorption of nitric acid onto alkaline aerosol particles is the

Fig. 13. Measurement results from Ispra ŽIT04..

228

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

most likely explanation. Sulfate concentration data are not available from the Danish station Lille Valby, but it would appear likely that ammonium nitrate is a major component, together with ammonium sulfate. At SE02 and NO01 ammonium nitrate is only observed during some episodes. Generally nitrate is associated with other cations. This is in accordance with previous observations, which show that nitrate is often associated with sea-salt particles. The low concentrations at SE12 ŽAspvreten, Sweden. makes it difficult to draw firm conclusions, but it appears that ammoniarammonium concentrations are depleted in the formation of ammonium sulfate, and that any nitrate present is associated with other cations than ammonium. 4.2. Lead and cadmium elements associated with particulate matter Airborne particulate matter includes trace

components such as cadmium and lead elements. Measurements of heavy metals in aerosols are already included in the EMEP programme. Here we briefly summarise the EMEP findings of ambient concentrations of lead and cadmium in aerosols in Europe during 1998 ŽBerg and Hjellbrekke, 2000.. It should be noted that the maps presented in Figs. 14 and 15 are based on relatively few measurement points and give only a coarse picture of the concentration distribution. Fig. 14 presents the annual averages of Pb in air in 1998 integrated with a kriging method. The lowest concentrations Ž- 1 ng Pbrm3 . can be seen on Spitsbergen ŽNO42. and in Iceland. A region with concentrations between 8 and 12 ng Pbrm3 can be seen in central parts of Europe. Maximum concentrations have been measured at the Sloval stations with annual means close to 20 ng Pbrm3. Cadmium in aerosols is presented in Fig. 15. As for Pb the lowest concentrations Ž- 0.05 ng Cdrm3 . are reported for Spitsbergen and Ice-

Fig. 14. Annual average concentrations Žngrm3. of lead in aerosols Ž1998..

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

land. An increasing gradient can be seen southeastward, with the highest concentration maxima at Slovak sites.

229

chemical analyses. In the period between 1991 and 1995 the sampling was done with the ‘Gent’ stacked filter unit, PM 10 impactor inlet, followed by two Nucleopore filters in a NILU filter holder. The first Nucleopore filter Ž‘coarse’. has an 8-␮m nominal pore size. At a flow rate of 15᎐16 lrmin, this gives a 50% efficiency cut-off at 2 ␮m. The second filter is also a Nucleopore filter with a nominal pore size of 0.4. Annual averages for Birkenes are indicated in Table 3. There has been a clear reduction in the concentration of inorganic aerosol constituents, particularly the elements associated with dust and fly ash emissions from process industries and solid

4.3. Chemical speciation of PM10 measurements in Birkenes (Norway) In this section we present an outline of research measurements on particulate matter PM 10 concentration and chemical composition in Birkenes ŽNorway. from the early 1970s until recently. The sampling in the campaigns before 1991 was performed with high-volume samplers, whereas the chemical analyses were done with neutron activation, atomic absorption and other

Table 3 Overview of aerosol concentrations at Birkenes, 1973᎐1993 Spring 1973 Si Al Fe Ca nss Ca K nss K Zn Pb Minerals Na Cl Br I Sea-salt

433 114 96 80 65 138 123 29

Autumn 1973 323 85 95 74

1978᎐ 1979

1985᎐ 1986

1991, coarse

1991, fine

1992 coarse

1992 fine

1993 coarse

1993 fine

304 80 84

277 73 61

18 1019

15 11 908

115 58 35 47 37 37 27 2 1 561

68 14 14 11 7 36 32 6 4 257

131 64 41 47 35 39 27 1.5 1.10 620

46 14 13 12 8 24 20 4.2 3.20 195

66 26 17 31 13 29 12 1.7 0.70 280

33 10 10 11 4 28 22 6.4 3.50 148

350 380 5.3

273 258

116 33

329 316 1.03 0.2 730

118 17 2.55 0.8 165

464 600 1.29

179 84 2.68 0.72 310

᎐

1658

35 23 1090

403 60

478 256

568

858

1206

821

602

179

2560 2520 n.a. 960 3480

4260 4217 n.a. 1598 5815

3060 3031 n.a. 1148 4178

393 370

354 326 n.a. 133 459

1797 1787

378 339

1740 1725

147 518

2712 2702 n.a. 1017 3719

674 2461

142 481

653 2378

5429

8040

5907

80 1760

544 4699

90 1899

410 3231

80 2026

514 3349

2800 1040

6780 2081

3120 1221

5018 1787

2898 872

5037 1688

SO4 nss SO4 NO3 NH4 Secondary inorganic Black carbon Sum quantified components Measured Unexplained

508 566 4.7

1185

All concentrations are given in ngrm3. Inferred values in italics. In spring 1973 and Autumn 1973 the data are based on neutron activation data compiled by Semb Ž1978.; 1978᎐1979 data are from Pacyna et al. Ž1984b.; 1985᎐1986 data are from Amundsen et al. Ž1992.; data from 1991 onwards are from Maenhaut et al. Ž1993.. Žnss: non-sea salt..

230

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

fuel combustion. Emissions of trace elements have also been reduced, and the concentration of ammonium sulfate is reduced since 1980. It is interesting to discuss the results in terms of different contributions to the total mass. For this purpose the following components have been defined. The ratio among the elements Si, Al, Fe and Ca is not very different from the average composition of the Earth’s crust, but Fe and Ca may be somewhat higher due to emissions from iron and steel industries, and from cement works. Potassium occurs in much higher concentrations than expected, and is particularly enriched in the PM 2.5 fraction. The marine aerosol consists mainly of Na and Cl, with minor amounts of Mg, Ca, K, and SO42y. The samples are always depleted with respect to Cl, this is particularly the case for the PM 2.5 fractions. Bromine is enriched in the fine fraction, so is iodine. Neither make any significant contribution to the particulate mass. Only sulfur has been determined by PIXE.

Other measurements have shown that sulfate is mainly present as ammonium sulfate, and that nitrate at Birkenes occurs in the coarse particle fraction, in association with sea-salt particles. The contribution of ammonium sulfate to the PM 2.5 mass is very substantial. Only elementary carbon has been determined with an optical method, which is not an absolute method. Nevertheless, the measured average concentration at 0.5᎐0.6 ␮grm3. The difference between the calculated contributions from the components which have been quantified on the basis of chemical analyses, and the particulate mass determined by weighing of the sample filters, have been labelled ‘undefined’. Both primary particulate matter and secondary organic particulate matter also contribute to the total mass of the PM 2.5 fraction. On the basis of measurements in urban areas, organic aerosol concentrations should be expected to be at least twice the concentration of elementary carbon. It is important to quantify these contributions

Fig. 15. Annual average concentrations Žngrm3. of cadmium in aerosols Ž1998..

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

231

Fig. 16. Measured PM 10 mass concentrations at Birkenes for the period February 1991᎐February 1996, together with inferred masses of specific components from chemical analyses.

properly. Primary particulate matter of biological origin Žbioaerosols, organic debris. may be a significant contributor to the 2.5- PM - 10 fraction. The contribution from minerals, determined on the basis of Si, Al, Fe and Ca, is concentrated in short time periods. The high concentrations in May᎐June 1992 are associated with air passing across southern Sweden from the Baltic coun-

tries, possibly implicating emission sources in the Narva᎐Leningrad region. It is expected that the minerals come from anthropogenic emissions, combustion of solid fuels, cement production, iron and steel industry Žsee Figs. 16 and 17.. Sea-salt is also a main contributor to the coarse particle fraction. Chloride is depleted relative to sodium in most of the samples, particularly in the fine particle fraction. Part of the depletion is

Fig. 17. Aerosol coarse fraction concentration during 1991 for a number of inorganic compounds.

232

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

caused by reaction of nitric acid with the sea-salt particles, displacing hydrochloric acid. However, these data are not suitable for establishing a quantitative relationship. A number of elements are ‘enriched’ in aerosols relative to their abundance in the earth’s crust or in seawater. These are mainly found in the fine fraction, and include elements such as lead, cadmium, zinc, arsenic, antimony, selenium, bromine and iodine. Neither of these contribute significantly to the particle mass, and, since leadfree petrol has become common, none are no longer valuable as tracers of particular emission sources either. The main constituents in the fine fractions appears to be ammonium sulfate, elementary carbon and organic compounds. Clearly, good analytical methods for the two latter components are desirable for the full quantification of the fine particle fraction, although minerals and sea-salts are also of some importance. The measurements of the fine particle mass are quite consistent with time, with good correlation between the total mass and the identified components. A number of tentative conclusions can be reached from the above studies. The Ghent stacked filter unit is useful for quantifying PM 10 and PM 2.5 , and can provide samples for subsequent chemical analyses. The low air volumes make the weighing somewhat demanding, and the membrane filters are also subject to electrostatic charging. On the other hand, these filters do not need the careful ‘conditioning’ which has to be carried out when weighing glass fibre filters. In addition, secondary inorganic particulate matter is the main constituent in the PM 2.5 fraction at background sites, explaining 30᎐50% of the total mass. The second most important component in the PM 2.5 fraction is black carbon or soot, with associated organic compounds from combustion processes. Quantification of this component, preferably with more detailed characterisation of the organic fraction, is the key to understanding the composition and properties of PM 2.5 , including the formation of secondary organic aerosols. Sea-salt components and minerals are of minor importance for the PM 2.5 fraction but account for a major part of the PM 10 ᎐PM 2.5 . These

are important in connection with deposition of sea-salt ions, and alkaline base cations.

5. Conclusions Data from epidemiological and toxicological studies demonstrate associations between ambient particulate concentrations and resulting health effects even though it is not yet clear the factors Žmass, chemical composition, numberrsize distribution. that are responsible for these effects. Therefore there is an urgent need for more detailed investigations, and in particular the understanding of the regional compositionrsize distribution characteristics of particulate matter and its relation to human health in the European scale. The European Monitoring and Evaluation Programme ŽEMEP. aims to provide a basis for a quantitative assessment of the long-range transported aerosol component, and their rural concentration levels. Although this information is already available for the secondary inorganic aerosol particulate matter, which is a significant part of the long-range transported aerosol, very little is known about the European-wide distribution of other components of particulate matter. In addition, measurements of aerosol particulate mass have mainly been made in urban environments urging the need for regional studies in Europe. It is apparent that harmonised measurements and sampling of PM 10 and more detailed chemical analyses of the particulate matter contribution to PM 10 and PM 2.5 are required to support the formulation of environmental policies to reduce ambient PM concentrations. The chemical analyses need to include elementary and organic carbon ŽECrOC., as well as other parameters that may help to identify primary and secondary organic matter, and natural aerosol components. Furthermore, the preliminary emission estimates needs to be refined, particularly with respect to chemical composition of the primary particles from different source categories. The above research needs are also main priorities in the EMEP programme in the coming years.

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

References Aas W, Hjellbrekke A-G, Semb A, Schaug J. Data quality 1997, quality assurance, and field comparisons. EMEPrCCC Report 6r99, 1999. Allen G, Sioutas C, Koutrakis P, Reiss R, Lurmann FW, Roberts PT. Evaluation of the TEOM method for measurement of ambient particulate mass in urban areas. J Air Waste Manage Assoc 1997;47:682᎐689. Alpert DJ, Hopke PK. A determination of the sources of airborne particles collected during the regional air pollution study. Atmos Environ 1981;15:675᎐687. Amundsen CE, Hanssen JE, Semb A, Steinnes E. Long-range transport of trace elements to southern Norway. Atmos Environ 1992;26A:1309᎐1324. Andreae MO, Crutzen PJ. Atmospheric aerosols: biogeochemical sources and role in atmospheric chemistry. Science 1997;276:1052᎐1058. Baedecker PA, Reddy MM, Reimann KJ, Sciammarella CA. Effects of acidic deposition on the erosion of carbonate stone ᎏ experimental results from the U.S. National Acid Precipitation Assessment Program ŽNAPAP.. Atmos Environ 1992;26B:147᎐158. Barrett K, Schaug J, Bartonova A, Semb A, Hjellbrekke A-G, Hanssen JE. A contribution from CCC to the reevaluation of the observed trends in sulphur and nitrogen in Europe 1978᎐1998. Input for further evaluation by the national laboratories and for use in the TFMM assessment work. CCC Report 7r2000, 2000. Berdowski JJM, Mulder W, Veldt C, Visschedijk AJH, Zandveld PYJ. Particulate emissions ŽPM10-PM2.5-PM0.1. in Europe in 1990 and 1993. Bilthoven: RIVM, 1998. Berg T, Hjellbrekke A-G. Heavy metals and POPs within the ECE region 1998. Kjeller, Norwegian Institute for Air Research ŽEMEPrCCC Report 7r00., 2000. Birch ME, Cary RA. Elemental carbon-based method for monitoring occupational exposures to particulate diesel exhaust. Aerosol Sci Technol 1996;25:221᎐241. Buijsman E, van Elzakker B. Meetaktiviteiten in 1996 in het kader van het Landelijk Meetnet Luchtkwaliteit. Bilthoven ŽRIVM Report 723101 022., 1996. Cachier H, Bremond M-P, Buat-Menard P. Determination of ´ atmospheric soot carbon with a simple thermal method. Tellus 1989;41B:379᎐390. CEN. Air quality ᎏ determination of the PM 10 fraction of suspended particulate matter ᎏ reference method and field test procedure to demonstrate reference equivalence of measurement methods. European Committee for Standardization ŽEuropean Standard EN 12341., 1998. Charlson RJ, Schwartz SE, Hales JM, Cess RD, Coakley Jr. JA, Hansen JE, Hofmann DJ. Climate forcing by anthropogenic aerosols. Science 1992;255:423᎐430. Chow JC, Watson JG, Pritchett LC, Pierson WR, Frazier CA, Purcell RG. The DRI thermalroptical reflectance carbon analysis system: description, evaluation and application in U.S. air quality studies. Atmos Environ 1993;27A: 1185᎐1201.

233

Clarke AD. Atmospheric nuclei in the remote free-troposphere. J Atmos Chem 1992;14:479᎐488. Covert DS, Kapustin VN, Quinn PK, Bates TS. New particle formation in the marine boundary layer. J Geophys Res 1992;97:20581᎐20587. Eliassen A, Saltbones J. Modelling of long range transport of sulphur over Europe: a two year model run and some model experiments. Atmos Environ 1983;17:1457᎐1473. EMEP-WMO Workshop on strategies for monitoring of regional air pollution in relation to the need within EMEP, GAW and other international bodies. Aspenas ¨ Herrgard, ˚ Lerum, Sweden June 2᎐4, 1997. Kjeller ŽEMEPrCCC-Report 10r97., 1997. EMEPrCCC. in: Hanssen JE, Schaug J, Semb A, editors. EMEP manual for sampling and chemical analysis ŽEMEPrCCC Report 1r95., 1996. EMEP transboundary acidification and eutrophication in Europe. CCC & MSC-W report ŽEMEP 1r00., 2000a. EMEP transboundary photo-oxidants in Europe. CCC & MSC-W report ŽEMEP 2r00., 2000b. CCC note ŽEMEP 1r01., 2001. EMEP Manual for measurements of PM 10 and chemical speciation of aerosol particles, 2001. EMEP-WMO. Workshop on Fine Particles-Emissions, Modelling and Measurements. Proceedings Interlaken, 22᎐25 November 1999 Switzerland. Kjeller, Norwegian Institute for Air Research ŽEMEP Report 9r2000. 1999. EPA. Air quality criteria for particulate matter. North Carolina, U.S. Environmental Protection Agency ŽEPAr600r P-95r001., 1996a. EPA. Exposure factors handbook. Technical Report. North Carolina, U.S. Environmental Protection Agency ŽEPAr600rP-95r002Ba., 1996b. EU Position Paper on Particles Ambient Air Pollution by Particulate Matter. Technical Working Group on Particles, 1998. European Commission. Ambient air pollution by particulate matter. Executive summary, 1996. European Commission. Proposal for a Council Directive relating to limit values for sulphur dioxide, oxides of nitrogen, particulate matter and lead in ambient air. Luxembourg, Office for Official Publications of the European Communities ŽCOMŽ97. 500., 1997. Finlayson-Pitts BJ, Pitts JN. Atmospheric chemistry: fundamentals and experimental techniques. Wiley, 1986. Hedin LO, Granat L, Likens GE, Buishand TA, Galloway JN, Butler TN, Rodhe H. Steep declines in atmospheric base cations in regions of Europe and North America. Nature 1994;367:351᎐354. Hegg DA, Ferek RJ, Hobbs PV. Light scattering and cloud condensation nucleus activity of sulfate aerosol measured over the northeast Atlantic Ocean. J Geophys Res 1993;98:14887᎐14894. Heintzenberg J, Muller K, Birmili W, Spindler G, Wiedensoh¨ ler A. Mass-related aerosol properties over the Leipzig basin. J Geophys Res 1998;103:13125᎐13135.

234

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235

Heintzenberg J. Fine particles in the global troposphere ᎏ a review. Tellus 1989;41B:149᎐160. Hoek G, Forsberg B, Borowska M, Hlawiczka S, Vaskovi E, Welinder H, Branis M, Benes I, Kotesovec F, Hagen LE, Cyrys J, Jantunen M, Roemer W, Brunekreef B. Wintertime PM 10 and black smoke concentrations across Europe: results from the PEACE Study. Atmos Environ 1997;31:3609᎐3622. Hoffmann T, Odum JR, Bowman F, Collins D, Klockow D, Flagan RC, Seinfeld JH. Formation of organic aerosols from oxidation of biogenic hydrocarbons. J Atmos Chem 1997;26:189᎐222. Lannefors H, Hansson H-C, Granat L. Background aerosol composition in southern Sweden ᎏ fourteen micro and macro constituents measured in seven particle size intervals at one site during one year. Atmos Environ 1983;17:87᎐111. Larssen S, Sluyter R. Criteria for EUROAIRNET. The EEA Air Quality Monitoring and Information Network Copenhagen, European Environment Agency ŽTechnical report 12., 1999. Larssen S, Lazaridis M. EUROAIRNET site selection 1998. Copenhagen, European Environment Agency ŽTechnical report 16., 1998. Lazaridis M, Semb A, Hov Ø. Long-range transport of aerosol particles. A literature review. Kjeller, Norwegian Institute for Air Research ŽEMEPrCCC-Report 8r99., 1999. Ligocki MP, Salmon LG, Fall T, Jones MC, Nazaroff WW, Cass GR. Characteristics of airborne particles inside southern California museums. Atmos Environ 1993;27A:697᎐711. Liousse C, Cachier H, Jennings SG. Optical and thermal measurements of black carbon aerosol in different environments. Atmos Environ 1993;27A:1203᎐1211. Maenhaut W, Francois F, Cafmeyer J. First Research Coordination Meeting of the International Atomic Energy Agency ŽIAEA.: Co-ordinated Research Programme on Applied Research on Air Pollution using Nuclear-Related Analytical Techniques. Vienna, Austria, 20 March᎐2 April. Proceedings, 1993. McMurry PH. A review of atmospheric aerosol measurements. Atmos Environ 2000;34:1959᎐1999. Pakkanen TA, Hillamo RE, Keronen P, Maenhaut W, Ducastel G, Pacyna JM. Sources and physico-chemical characteristics of the atmospheric aerosol in Southern Norway. Atmos Environ 1996;30:1391᎐1405. Pacyna JM, Semb A, Hanssen JE. Emission and long range transport of trace elements in Europe. Tellus 1984a; 36B:163᎐178. Pacyna JM, Ottar B, Hanssen JE, Kemp K. The chemical composition of aerosols measured in Southern Scandinavia. Lillestrøm ŽNILU OR 66r84., 1984b. Pacyna JM, Larssen S, Semb S. European survey for NO x emissions with emphasis on Eastern Europe. Atmos Environ 1991;25A:425᎐439. Patashnik H. Letter to the editor. J Air Waste Manage Assoc 1998;48:195᎐197. Patashnik H, Rupprecht EG. Continuous PM 10 measure-

ments using the tapered element oscillating microbalance. J Air Waste Manage Assoc 1991;41:1079᎐1083. Peirson DH, Cawse PA, Salmon L, Cambray RS. Trace elements in the atmospheric environment. Nature 1973;241: 252᎐256. Pelley J. Is coastal eutrophication out of control? Environ Sci Technol 1998;32:462A᎐466A. Pope CA, Thun MJ, Namboodiri MM, Dockery DW, Evans JS, Speizer FE, Heath Jr. DW. Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J. Respir. Crit. Care Med. 1995;151:669᎐674. Rodriguez S, Querol X, Alastuey A, Kallos G, Kakaliagou O. Saharan dust contributions to PM 10 and TSP levels in southern and eastern Spain. Atmos Environ 2001;35: 2433᎐2447. Rogge WF, Mazurek MA, Hildemann LM, Cass GR, Simoneit BRT. Quantification of urban organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos Environ 1993;27:1309᎐1330. Schlesinger RB. Toxicological evidence for health effects from inhaled particulate pollution: does is support the human experinece? Inhal Toxicol 1995;7:99᎐109. Schwartz J, Dockery DW. Increased mortality in Philadelphia associated with daily air pollution concentrations. Am Rev Respir Dis 1992;145:600᎐604. Schwartz J, Dockery DW, Neas LM. Is daily mortality associated specifically with fine particles? J Air Waste Manage Assoc 1996;46:927᎐939. Seinfeld J, Pandis S. Atmospheric chemistry and physics of air pollution: from air pollution to climate change. New York: Wiley, 1998. Semb A. Deposition of trace elements from the atmosphere in ˚ Norway. ŽSNSF FR 13r78., Norway. SNSF Projeckt, As, 1978. Semb A, Hanssen JE, Francois F, Maenhaut W, Pacyna JM. Long range transport and deposition of mineral matter as a source for base cations. Water Air Soil Pollut 1995;85:1933᎐1940. Semb A, Bartonova A, Schaug J, Lukeville A, Tørseth K. Pilot ¨ measurements of nitrogen containing species in air. Kjeller, Norwegian Institute for Air Research ŽEMEPrCCC-Report 5r98., 1998. Tarrason L, Tsyro S. Long-range transport of fine secondary particles, as presently estimated by the EMEP Lagrangian model. Oslo, DNMI ŽEMEPrMSC-W Note 2r98. ŽResearch reportrNorwegian Meteorological Institute no. 69., 1998. Trijonis JC, Malm WC, Pitchford M, White WH. Visibility: existing and historical conditions ᎏ causes and effects. In: Irving PM, editor, 1991. Turpin BJ, Saxena P, Andrews E. Measuring and simulating particulate organics in the atmosphere: problems and prospects. Atmos Environ 2000;34:2983᎐3013. Quality of Urban Air Review Group. Airborne particulate matter in the United Kingdom. UK, 1996.

M. Lazaridis et al. r The Science of the Total En¨ ironment 285 (2002) 209᎐235 Wedding JB, Weigand MA. An automatic particle sampler with beta gauging. J Air Waste Manage Assoc 1993;43:475᎐479. WHO. Particulate matter ŽPM 10 and PM 2.5 .. Results of intercomparison studies. Report on a WHO Workshop, Berlin 3᎐5 September 1998. Copenhagen, WHO Regional Office for Europe ŽEURrICPrEHB1040102, E62010, 10-13., 1999. Wilson WE, Suh HH. Fine particles and coarse particles: Concentration relationships relevant to epidemiologic studies. J Air Waste Manage Assoc 1997;47:1238᎐1249.

235

Van der Zee SK, Hoek G, Harssema H, Brunekreef B. Characterization of particulate air pollution in urban and nonurban areas in the Netherlands. Atmos Environ 1998;32:3717᎐3729. Zappoli S, Andracchio A, Fuzzi S, Facchini MC, Gelencser A, Kiss G, Krivacsy Z, Molnar A, Meszaros E, Hansson H-C, Rosman K, Zebuhr Y. Inorganic, organic and macromolecular components of fine aerosol in different areas of Europe in relation to their water solubility. Atmos Environ 1999;33:2733᎐2743.