1

Development and implementation of a new SOA scheme in CHIMERE -Model evaluation against organic and elemental carbon observations

1Bertrand BESSAGNET, 2Laurent MENUT, 3Gabriele CURCI, 4Alma HODZIC, 5Bruno

GUILLAUME, 5Catherine LIOUSSE, 6Sophie MOUKHTAR, 7Betty PUN, 7Christian

SEIGNEUR, 8Michael SCHULZ

(1) INERIS, Institut National de l’Environnement Industriel et des Risques, Parc

technologique ALATA, 60550 Verneuil en Halatte, France.

(2) Institut P.-S. Laplace, Laboratoire de Meteorologie Dynamique, Ecole Polytechnique,

F-91128 Palaiseau, France.

(3) CETEMPS, Universita degli Studi dell’Aquila, via Vetoio, 67010 Coppito - L’Aquila,

Italy.

(4) NCAR, National Center for Atmospheric Research, 3450 Mitchell Lane, 80301, CO, USA.

(5) LA/OMP, Laboratoire d’Aerologie / Observatoire Midi-Pyrenees, 14, avenue Edouard

Belin, 31400 Toulouse, France.

(6) Centre for Atmospheric Chemistry - York University - 4700 Keele Street, Toronto, Canada.

(7) Atmospheric & Environmental Research, 2682 Bishop Drive, Suite 120, San Ramon, CA

94583, USA.

(8) Laboratoire des Sciences du Climat et de l’Environnement, IPSL/CEA-CNRS-UVSQ,

F-91190 Gif-sur-Yvette, France.

Short title: CARBONACEOUS SPECIES IN CHIMERE1

2

Abstract.2

The current state of the art of chemistry-transport modeling shows severe limitations for3

Secondary Organics Aerosols (SOA) concentrations estimation. This often implies difficulties to4

forecast or investigate polluted situations since important sources and their corresponding chemistry5

are not taken into account. In this study, a new and complete SOA chemistry scheme was implemented6

in the CHIMERE chemistry-transport model. Moreover, in order to improve the emission sources, the7

MEGAN biogenic emissions inventory was implemented. Hourly simulations were performed over the8

entire year of 2003 in Western Europe. The model results were clearly improved by implementing the9

new SOA scheme. In this study, a clear underestimation of OC concentrations was diagnosed during10

winter. A possible explanation could be missing (but observed) wood burning emissions in Portugal,11

Italy, Slovakia and Hungary as suggested by other modeling works. In addition, this work suggests that12

during the higher fire emission periods, fires can be the dominant source of primary organic carbon over13

the Mediterranean Basin. The contribution of SOA from fire emissions is low. This study highlights14

more precisely reasons of discrepancies between observations and modeling choices. We show that15

isoprene chemistry has a strong impact on SOA formation when using the current available kinetic16

schemes. Such results could explain large underestimates of OC concentrations in the southern Europe17

when this specific chemistry is not accounted for.18

3

1. Introduction19

Particulate matter (PM) pollution control is one of the main challenge highlighted20

by the Thematic Strategy on Air Pollution (CAFE as Clean Air For Europe), adopted by21

the European Commission in October 2005, under its 6th Environmental Action Program22

(Decision No 1600/2002/EC of the European Parliament and of the Council of July 22, 200223

laying down the Sixth Community Environment Action Programme). The CAFE strategy24

states that particulate matter (and especially fine particles with diameter smaller than 2.5 µm25

- PM2.5) is responsible today for an average reduction of life expectancy of about 8 months26

in Europe: recent epidemiological studies highlight the role of the smallest part of these27

particles on health [Schlesinger et al., 2006; Lee et al., 2007; Heinrich and Slama, 2007]. The28

fine particles are composed of a large fraction of organic (OC) and elemental carbon (EC)29

([Putaud et al., 2004]). According to [Van Dingenen et al., 2004], 70% to 80% of particles30

number consists mainly of carbonaceous material. This carbonaceous material is injected into31

the atmosphere by diffusion and mixing of surface emissions or by direct injection of biomass32

burning. They can also be formed by chemical reactions of Volatile Organic Compounds33

(VOC) in the atmosphere, the so called Secondary Organic Aerosols (SOA) (see [Kroll and34

Seinfeld, 2008] for a complete review of the SOA chemistry in gas and aqueous phases). This35

includes the large contribution of terpenes and isoprene ([Surratt et al., 2006]) emitted by36

the vegetation. While in winter biomass burning emissions are the main source with sizable37

additional contribution from fossil fuel combustion, the SOA originate mainly from non fossil38

sources in summer. Based on the CARBOSOL measurements, [Gelencser et al., 2007] showed39

4

that non fossil sources represent 63-76% of the Total Carbon (hereafter TC). [Pio et al.,40

2007] shows that 50-80% of OC is water soluble and suggest that OC has to be considered41

in discussing the role of clouds on climate over Europe. Humic-like substances (HULIS) are42

contributors of water soluble compounds (3-7%) and are directly emitted by biomass burning43

and probably formed by chemical reactions in the atmosphere ([Mayol-Bracero et al., 2002;44

Graber and Rudich, 2006; Lukacs et al., 2007; Schmidl et al., 2008]). A chemical mechanism45

in aqueous phase through the photo-oxidation of methylglyoxal is proposed by [Altieri et al.,46

2008] for HULIS formation. According [Legrand and Puxbaum, 2007] and references therein,47

it seems that most OC is contained in oligomeric or polymeric matter. OC emitted by the48

decomposition of vegetative debris is an other source of the organic coarse size fraction49

([Puxbaum and Tenze-Kunit, 2003]), [Bauer et al., 2002] found a contribution of 6% in a50

remote site.51

In the framework of the CARBOSOL campaign, and in a model context, the modeled52

EC and OC concentrations fields of the EMEP model were compared to measurements.53

OC is generally underpredicted in most of the sites certainly due to missing wood burning54

contributions. The model underpredicts TC in the southern Europe mainly due to a lack55

of SOA ([Simpson et al., 2007]). For EC, the largest uncertainties probably lie in EC56

emissions from residential wood/fossil combustion possibilly associated with both emission57

factors and spatial and temporal variation ([Tsyro et al., 2007]). During a specific modeling58

intercomparison exercize, a potential large underestimate of EC components has been recently59

reported in [Stern et al., 2008] in Eastern Germany, due to underestimated sources in Eastern60

Europe. Over the United-States, the role of isoprene in secondary organic aerosol formation61

5

([Dommen et al., 2006; Kroll et al., 2006]) has been extensively studied by [van Donkelaar62

et al., 2007] and [Zhang et al., 2007] that suggests a high sensitivity to the values of the63

enthalpy of vaporization used in models.64

The CHIMERE model was extensively validated on PM10, sulfate, nitrate and ammonium65

components in [Bessagnet et al., 2004, 2005; Vautard et al., 2005; Hodzic et al., 2005]. In this66

work, the SOA scheme implemented in CHIMERE is presented, a new emission inventory67

for biogenic VOC emissions has been implemented. A first evaluation of the CHIMERE68

model is proposed for carbonaceous species by using OC and EC observations from the69

CARBOSOL and EMEP available data for the year 2003. This year was characterized by70

huge fire emissions in the southern Europe taken into account in this study.71

2. The chemistry transport model CHIMERE72

2.1. General description73

Given a set of NOx, SOx, NH3, PM, VOC’s and CO emissions, the CHIMERE model74

calculates the concentrations of 44 gas-phase and aerosol species. In this study, a version75

of CHIMERE for a domain covering the western Europe is used [Schmidt et al., 2001;76

Bessagnet et al., 2004]: from 14oW to 28oE in longitude and from 35oN to 58.2oN in77

latitude, with a constant horizontal resolution of 0.4o× 0.4o. The vertical grid contains78

15 layers from surface to 500 hPa. The dynamics and gas-phase parts of the model are79

described in [Schmidt et al., 2001], and improvements have successively been brought80

([Vautard et al., 2003, 2005]). The model documentation can be found on the web server81

6

http://euler.lmd.polytechnique.fr/chimere. For both ozone and PM10, the model has undergone82

extensive modeled aerosols intercomparisons at European and city scales ([Vautard et al.,83

2007; Van Loon et al., 2007; Schaap et al., 2007]).84

The CHIMERE model is hourly driven by the meteorological model MM5 for the85

dynamical parameters (wind, temperature, humidity, pressure). In order to be consistent86

with the PREVAIR operational forecast system using MM5 and CHIMERE, the current87

MM5 configuration ([Dudhia, 1993]) was chosen to be the same than the one used and88

validated during the last three years for the daily forecast, [Honore et al., 2008]. The MM589

vertical grid contains 32 levels ranging from surface to 10hPa. The horizontal resolution is90

54km over a domain encompassing the european CHIMERE domain. The meteorological91

boundary conditions as well as the nudging is performed with the six-hourly ECMWF analysis92

meteorological fields.93

The aerosol module is that described in [Bessagnet et al., 2004]. Anthropogenic gas94

emissions are taken from the Co-operative Programme for Monitoring and Evaluation of the95

Long-range Transmission of Air pollutants in Europe - EMEP - ([Vestreng, 2003]). Three96

particulate model species are considered: PPM (primary particle material) that contains only97

mineral dust from anthropogenic sources (EMEP inventory) assumed to be coarse particles,98

primary OC and EC assumed to be in the fine mode. OC and EC emissions are from a specific99

inventory discussed in the next section. Calculation of model species emissions that can be100

used by the CHIMERE CTM, is made in several steps. The spatial emission distribution from101

the original grid to the CHIMERE grid is performed using an intermediate fine grid with a102

1km resolution (GLCF dataset, [Hansen et al., 2000]). Soil types being known on the fine103

7

grid allows for a better apportionment of the emissions according to urban, rural, maritime104

and continental areas. Standard time variation profiles are applied to get hourly emission105

from annual data, as required by the model. The other modeled species are sulfates, nitrates,106

ammonium, secondary organic aerosols, sea-salt (considered as inert here) and dust. The107

particle size distribution ranges from about 40 nm to 10 µm and are distributed into 8 bins.108

The 8 bins used are defined between the following intervals: 0.039, 0.078, 0.156, 0.312, 0.625,109

1.25, 2.5, 5, 10 µm. The gas - particle partitioning of the ensemble Sulfate/Nitrate/Ammonium110

is treated by the code ISORROPIA ([Nenes et al., 1998]) implemented in CHIMERE. For the111

main gas and aerosols, boundary conditions are issued from a 5 years climatology (2001-2005)112

of the global model LMDzT-INCA. For aerosol boundary conditions, only elemental and113

organic carbon, desert dust and sulfate are taken into account. The LMDzT-INCA model114

contains a multi-modal aerosol distribution ([Textor et al., 2006; Schulz et al., 2006]). Organic115

and elemental carbon are described as belonging to a soluble and insoluble accumulation116

mode, where ageing processes transfer constantly insoluble matter into the soluble aerosol117

mode. The emissions are described by [Dentener et al., 2006] and have been used for the118

AeroCom model intercomparison experiment B.119

2.2. OC and EC emissions over Europe120

2.2.1. Primary OC and EC Emission inventories over Europe A particular effort has121

been put on primary OC (OCp) and EC emission inventories emitted by fossil fuel and biofuel122

combustion (traffic, industry, residential sources), which are key elements in EC and OC123

modeling. These emissions, which are traditionally obtained by bottom-up approaches are still124

8

hampered by severe uncertainties resulting from large differences in the choice of emission125

factors. Two major different approaches for deriving global fossil fuel and biofuel EC and126

OCp emission inventories are currently available, [Junker and Liousse, 2008] and [Bond et al.,127

2004], the main difference being in technology differentiations. Thus, [Bond et al., 2004]128

consider for each fossil fuel a detailed list of combustion technologies and emission controls129

with associated emission factors (EF), while in [Junker and Liousse, 2008], for each fuel, two130

parameters characterize the emissions. First, the activity sector (domestic, industrial, traffic)131

and second, the national level of development (with three levels: developed, semi-developed132

and developing countries), based on gross national incomes taken as a proxy reflecting133

technology and emission control levels. In each of these methods, the part of uncontrolled134

emissions is differently handled which is the origin of their largest differences. Higher135

estimated emissions may be found for the major fuels (coal, diesel, peat, lignite, coke) in136

[Junker and Liousse, 2008] than in [Bond et al., 2004] with more controlled emissions in [Bond137

et al., 2004] than in [Junker and Liousse, 2008]. Harmonization between these two inventory138

types is in progress at a global scale. At the European scale, detailed inventories are given139

for EC and OCp emissions in [Schaap et al., 2004] (after [Bond et al., 2004] methodology140

and emission factors) and in [Guillaume et al., 2007, 2008]. A technology-splitting approach141

has been adopted with the same considerations of emission controls and details as in [Schaap142

et al., 2004] and in [Bond et al., 2004], while keeping EF values estimated from [Junker and143

Liousse, 2008]. With such a configuration, both newer combustor types and ”large emitters”144

that still exist in Europe, are taken into account. The IIASA fuel dataset ([Klimont et al.,145

2002]) are used, covering 35 European countries excluding Russia and Ukraine and including146

9

informations on controlled and uncontrolled fractions of fuel use which depend on emission147

control set up. Let us note that waste burning is not included. Emission factors (EF) for EC148

and OCp are obtained from [Junker and Liousse, 2008], using a proxy when EF values are not149

welknown. Variation of EC/total particulate matter ratio is the usual proxy used to derive150

unknown EF values. This proxy is based on relationships between EC/OCp and CO/CO2151

ratios. It is also interesting to note that an improved spatial distribution of EC and OCp152

emissions than in previous inventories is used with traffic and domestic emissions scaled on153

rural and urban population densities, but industrial emissions spatialized according to their real154

geographical positions and magnitudes. A comparison between this inventory and [Schaap155

et al., 2004] is proposed in [Guillaume et al., 2007]. In both inventories, EC from brown156

coal and hard coal fuels dominates while slowly replaced by diesel; wood relative importance157

is comparable. Fuel consumption is generally in agreement. Main differences occur in the158

relative importance of controlled and uncontrolled fractions of fuel use and in brown and159

hard coal emission factors selected for industrial and domestic sectors. Emission factors by160

[Schaap et al., 2004], based on [Bond et al., 2004] values are smaller than in [Guillaume161

et al., 2007]. Industrial emissions (especially power plant emissions) are more controlled in162

[Schaap et al., 2004] than in the inventory used in this work. These differences are particularly163

important for Poland emissions, much higher in this inventory than in [Schaap et al., 2004].164

Same differences exist for domestic emissions while traffic emissions are comparable. Finally,165

this European EC inventory is about 1.5 times higher than in [Schaap et al., 2004], while in166

agreement with the European zoom of [Junker and Liousse, 2008] inventory. Both inventories167

have been tested in ORISAM-TM4 global transport model [Guillaume et al., 2007, 2008].168

10

Following these results, EC and OCp European emission inventories for the year 2000 and169

built with a 25kmx25km resolution, have been selected in our study. These emissions are170

extrapolated to the CHIMERE resolution.171

2.2.2. Fire emissions In the past decade, wildfires have devastated vast areas of forest172

and agricultural lands across Europe. In 2003 alone, more than 650,000 hectares of forest area173

and about 45,000 ha of agricultural land have been destroyed in Europe, which released into174

the atmosphere large amount of smoke particles and trace gases (such as CO, CO2, VOC’s,175

NO, NO2, etc.). The VOC speciation of [EPA, 1993] was used. The 2003 wildfire emissions176

contributed significantly to the enhancement of carbonaceous aerosol concentrations and177

perturbations of the surface radiative balance. In order to accurately assess the effects of178

wildfires on the atmospheric chemistry and radiative budget, the amount of emitted species179

needs to be quantified. In this study, and in the same way than in [Hodzic et al., 2007],180

daily wildfires emissions of particulate matter and trace gases together with their geographic181

location were estimated based on satellite information including (i) the location and date of182

the fire event, (ii) the area burned, (iii) the fuel loading factors (mass of biomass per area),183

(iv) the fraction of biomass fuel burned, and (v) the emission factors. These parameters have184

been determined by combining data available from several satellite products according to the185

methodology described in [Wiedinmyer et al., 2006]. The VOC speciation of [EPA, 1993] was186

used. A detailed description of the emission dataset is given by [Hodzic et al., 2007].187

11

2.3. SOA modeling in CHIMERE188

The complete chemical scheme implemented in CHIMERE includes biogenic and189

anthropogenic precursors (Table 1). Biogenic precursors include API (α-pinene and190

sabinene), BPI (β-pinene and δ3-carene), LIM (limonene), TPO (myrcene and ocimene)191

and ISO (isoprene). Anthropogenic precursors include TOL (benzene, toluene and other192

mono-substituted aromatics), TMB (Trimethylbenzene and other poly-substituted aromatics),193

and NC4H10 (higher alkanes). SOA formation is represented according to a single-step194

oxidation of the relevant precursors and gas-particle partitioning of the condensable oxidation195

products. The gas-particle partitioning formulation has been described in detail by [Pun et al.,196

2006]. The overall approach consists in differentiating between hydrophilic SOA that are most197

likely to dissolve into aqueous inorganic particles and hydrophobic SOA that are most likely198

to absorb into organic particles. The dissolution of hydrophilic SOA is governed by Henry’s199

law whereas the absorption of hydrophobic particles is governed by Raoult’s law. The large200

number of condensable organic compounds is represented by a set of surrogate compounds201

that cover the range of physico-chemical properties relevant for aerosol formation, i.e., water202

solubility and acid dissociation for hydrophilic compounds and saturation vapor pressure for203

hydrophobic compounds. These surrogate compounds were selected by grouping identified204

particulate-phase molecular products with similar properties. The molecular weight of each205

surrogate compound is determined based on its structure and functional groups. The Henry’s206

law constant or the saturation vapor pressure of the surrogate species is derived from the207

average properties of the group. Other properties are estimated using the structure of each208

12

surrogate compound.209

The absorption process in CHIMERE is implemented as in [Bowman et al., 1997]. A210

dynamical approach is adopted to decribe the gas particle conversion.211

Ji =1

τi

(Gi −Geqi ) (1)

Ji (µg.m−3.s−1) is the absorption or desorption flux of species i, τi (s) is a characteristic212

time of the mass transfer that depends on particle size and the chemical properties of species i,213

Gi is the bulk gas-phase concentration of species i and Geqi is the gas-phase concentration of214

species i at thermodynamic equilibrium (i.e., at the surface of the particle). The equilibrium215

gas-phase concentrations are functions of the particle chemical composition, temperature and,216

for hydrophilic species, relative humidity, as described by [Pun et al., 2006].217

The base SOA module was tested against the smog chamber data of [Odum et al., 1997]218

for anthropogenic compounds and those of [Griffin et al., 1999] for biogenic compounds219

and was shown to satisfactorily reproduce SOA formation for those compounds [Pun et al.,220

2006]. Higher alkanes and isoprene were added to the original chemical mechanism of [Pun221

et al., 2006]. The formation of SOA from higher alkanes follows the formulation of [Zhang222

et al., 2007] for the stoichiometric SOA yield and it is assumed that the SOA species can be223

represented by a hydrophobic surrogate compound with a moderate saturation vapor pressure.224

The formation of SOA from the oxidation of isoprene by hydroxyl radicals is represented with225

two surrogate products and follows the formulation of [Kroll et al., 2006; Zhang et al., 2007].226Table 1.

13

2.4. Implementation of the MEGAN biogenic inventory227

The previous biogenic inventory used in CHIMERE was based on Simpson’s algorithms228

([Simpson et al., 1995; Moukhtar et al., 2005]). In CHIMERE, a strong and questionable229

assumption was the use of a unique forest composition for a given country. In order to230

bypass this problem, the Model of Emissions of Gases and Aerosols from Nature (MEGAN,231

[Guenther et al., 2006], v. 2.04) was implemented for this study in the CHIMERE model. It232

estimates emissions of Volatile Organic Compounds and NO from vegetation as :233

ERi = EFi × γi(T, PPFD, LAI)× ρi (2)

where ERi (µg.m−2.h−1) is the emissions rate of species i, EFi (µg.m−2.h−1) is an234

emission factor at canopy standard conditions, γi (unitless) is an emission activity factor that235

accounts for deviations from canopy standard conditions, and ρi is a factor that accounts236

for production/loss within canopy. The canopy standard conditions relevant for this study237

are defined as: air temperature (T) of 303 K, photosynthetic photon flux density (PPFD) of238

1500 µmol.m−2.s−1 at top of the canopy, leaf area index (LAI) of 5 m2.m−2 and a canopy239

with 80% mature, 10% growing and 10% old foliage. The MEGAN model parameterizes the240

bulk effect of changing environmental conditions using three time-dependent input variables241

specified at top of the canopy: temperature (T), radiation (PPFD), and foliage density (LAI).242

The production/loss term within canopy is assumed to be unity (ρ = 1). The equation can then243

be expanded as :244

14

ERi = EFi × γT,i × γPPFD × γLAI (3)

The MEGAN model provides input EF and LAI data over a global grid, herefater245

projected on the CHIMERE model grid. The current available choice for EF’s is restricted246

to following species: isoprene, α-pinene, β-pinene, myrcene, sabinene, limonene, δ3-carene,247

ocimene, and nitrogen oxide. EF’s are static and refer to years 2000-2001. They are obtained248

summing up over several plant functional types (e.g. broadleaf and needle trees, shrubs, etc...).249

LAI database is given as a monthly mean product derived from MODIS observations, referred250

to base year 2000. Hourly emissions are calculated using 2-m temperature and short-wave251

radiation from MM5 model. The optimal choice for this work is the 150 seconds resolution252

(≈5 Km) products proposed in the MEGAN inventory.253Figure 1.

Figure 1 shows the differences for July 2003 between the new inventory MEGAN254

implemented in CHIMERE and the former approach based on [Simpson et al., 1995]. Large255

differences everywhere in Europe are observed for Terpene emissions with lower emissions256

using the MEGAN algorithms. Isoprene emissions are higher with MEGAN in Poland, Spain,257

Italy and Portugal and lower in Greece, United-Kingdom and North Africa.258

3. Comparisons between model and observational data259

3.1. Observational data260

Two databases of measurements are used in this study:261

1. The CARBOSOL data as described in [Pio et al., 2007; Legrand and Puxbaum, 2007]).262

15

EC, OC and diacids C2 to C5 (glutaric, malic, succinic, oxalic, malonic and tartaric263

acids) chemical analyses are available on a weekly basis. These low molecular weight264

dicarboxylic acids are mainly formed in the atmosphere by oxidation of VOC’s, but they265

also have a primary origin ([Limbeck and Puxbaum, 1999; Plewka et al., 2006; Kleefeld266

et al., 2002; Kawamura and Yasui, 2005]). Analyses are carried out for particles smaller267

than 2.5 µm.268

2. The EMEP data issued from the EMEP 2002-2003 campaign ([Yttri et al., 2007]). A269

daily sample per week with only EC and OC analyses is available. Analyses were270

performed using the thermal-optical transmission (TOT) instrument from Sunset Lab271

Inc., operating according to a NIOSH derived temperature program, more details can be272

found in [Yttri et al., 2007]. The routine measurements at the EMEP sites provide daily273

average PM10 concentrations (particle diameter smaller than 10 µm)274

The coordinates and locations in Europe of all stations used in this study are described in275

Table 2 and Figure 2. For the comparisons to modeled concentrations, values are taken in276

the first model layer. This may induce a problem when the topography varies a lot or in277

mountainous regions: the measurements are not really representative of the entire modeled278

cell. This is the case for the Puy de Dome (PDD) and Schauinsland (SIL) stations that are279

the most elevated sites. In order to convert the Organic Material (OM) concentrations to the280

modeled Organic Carbon (OC), the relation proposed by [Turpin and Lim, 2001] was used281

with OM = 1.6×OC.282 Figure 2.

Table 2.

16

3.2. OC comparisons283

In Figure 3 and Table 3, the comparison of CHIMERE with OC measurements shows a284

systematic underestimation for each site. Temporal correlations for PDD, SIL, BE02, GB46285

and IE31 sites are in the range 0.75 - 0.91. For IT04, IT08, AVE, PT01, SK04 and KPZ286

sites a clear underestimate by the CHIMERE model is observed in wintertime. The same287

underestimate by the EMEP model highlighted by [Simpson et al., 2007] was attributed to288

missing wood burning sources in these countries. In summertime, SOA largely dominates289

the organic fraction in simulated results, in the range 90-95% for the two elevated sites PDD290

(France) and SIL (Germany). These findings are in line with [Gelencser et al., 2007] who291

find that at least 86% of OC could be attributed to SOA. CHIMERE gives surprisingly high292

SOA contributions to total OC in winter for PDD, 71% and SIL, 52% for January-March and293

October-December periods. This is consistent with [Gelencser et al., 2007] findings that show294

a high fraction of biogenic secondary organic carbon in winter, 72% at SIL and 69% at PDD295

for the same periods.296Figure 3.

Table 3. The yearly averaged SOA/OM ratio simulated by CHIMERE are mostly in the range297

30-80 % in Europe (Figure 4). The highest SOA/OM ratio are found over the Pyrenees and298

Massif Central in France and in Spain. Poland displays lower SOA contribution in Europe299

because of higher primary organic emissions from industries in that country as previously300

discussed. In the southern Europe, secondary organic aerosols produced by isoprene chemistry301

dominates the total SOA formation (Figure 5). [Henze and Seinfeld, 2006] found a large302

increase of the SOA global budget by a factor of two by accounting for isoprene chemistry in303

17

SOA formation.304 Figure 4.

Figure 5.The evaluation of a SOA scheme is actually impossible because total SOA measurements305

do not exist yet in Europe. However, we can evaluate the impact of adding a SOA scheme on306

PM10 simulations. Some EMEP background sites in Europe have been selected.307 Table 4.

As shown in Table 4, the implementation of the SOA scheme improves the temporal308

correlation at many background sites. Large improvements are observed in Spain and Slovenia309

where biogenic VOC precursor emissions are very large. In Italy, in IT04 (Ispra) temporal310

correlations are impaired, but largely improved in IT01 (Montelibretti). In the forest region311

where the station DE08 is located in Germany, a large improvement is also observed from312

0.58 to 0.70. In Spain, a clear improvement is obtained by adding the isoprene chemistry for313

SOA formation. In Slovakia the temporal correlations are not globally improved. The urban314

site of Perigueux (PERI), a small city on the western part of the Massif Central, is located in a315

strong biogenic precursors emission area in France. In this region the CHIMERE simulation316

exhibits a high contribution of secondary organic species (up to 80%) on the total organic317

carbon as shown in Figure 4. The temporal correlation is largely improved at this site when318

implementing the SOA scheme from 0.39 to 0.51. Considering only the summer period from319

May 1st to September 30, the correlation is improved from 0.24 to 0.63 at Perigueux.320 Table 5.

For the sites AVE, PDD and SIL, the temporal correlation between observed diacid321

(C2-C5) concentrations and modeled SOA is fairly good (Table 5). Such a result could explain322

the good agreement between OC measurements and the simulated OC for PDD and SIL as323

reported in Table 3. For AVE, a good correlation is observed between diacids measurements324

and model SOA and a poor correlation between OC measurements and simulated OC, that325

18

confirms a problem with primary organic carbon in the model certainly due to missing326

emissions.327

3.3. EC comparisons328

Figure 6 and Table 6 display the comparison of EC model results against observations at329

each site. As for OC concentrations, the model underpredict EC concentrations at AVE, PT01,330

IT01, IT04 and KPZ particularly in winter. The assumption made by [Simpson et al., 2007]331

about missing wood burning sources in the model could again explain this behavior. The332

model reproduces well the temporal evolution of EC concentrations at BE02, NL09, IE31 and333

GB46, these sites are very close to well documented sources and are located over flat areas.334

For Mace Head (IE31), EC concentrations given by the model are globally underestimated335

by a factor of two: 0.24 µg.m−3 vs. 0.13 µg.m−3 respectively for the observations and the336

model. The low temporal correlation at the two elevated sites (SIL and PDD) suggest that337

these sites could be influenced by long range transport of anthropogenic pollution. By using338

monthly climatologies, the model is not able to capture daily intercontinental fluxes that can339

influence EC concentrations at elevated sites. That could be the reason why correlations340

for OC concentrations are higher than for EC concentrations at elevated sites because OC341

has a more local secondary origin (from the oxidation of VOC precursors) than EC. EC342

concentrations at remote places are affected by long range transport from anthropic emission343

areas ([Kasper and Puxbaum, 1998; Hitzenberger et al., 1999; Guillaume et al., 2008; Kaiser344

et al., 2007; Jaffe et al., 2003]) and discrepancies in meteorological calculation can largely345

impair model concentration results.346Figure 6.

Table 6.

19

3.4. Potential impact of forest fires347

Fire emissions can be an important contributor to OC concentrations as shown by348

[Langmann et al., 2008]. However, wild fire emissions do not contribute much to the349

atmospheric EC concentrations ([Tsyro et al., 2007]) on a yearly average. The impact of350

fire emissions depends on fire buoyancy and in models, sensitivity to fires depends on the351

altitude of their release on the vertical grid ([Hodzic et al., 2006]). Figure 7 shows the impact352

of adding forest fire respectively on primary and secondary organic material estimated by353

CHIMERE on August 1st - 15, 2003. During this period, intense fires were recorded around354

the Mediterranean basin and in Portugal. CHIMERE estimates give a large contribution of355

these fires to the primary organic concentrations in the South West of Europe with more than356

90 % in Portugal and often more than 50 % over the Mediterranean Sea. Moreover, fires357

emit volatile organic precursors that can be oxidized and form SOA. In Figure 7[bottom], the358

model gives a limited impact of secondary organic material originating from biomass burning359

related VOC’s, less than 10 % during the more intense fire period of the year 2003. The360

modeling results provide insight into a key question in [Gelencser et al., 2007] concerning361

the contribution of secondary organic carbon from VOC emitted by vegetation and biomass362

burning, because analysis methods cannot separate the two contributions.363 Figure 7.

4. Sources of uncertainties in SOA modeling364

From the measurements to the modeled scheme definition, the modeling of SOA includes365

a lot of uncertainties. Recently, the gap between laboratory studies (mandatory to develop366

20

more realistic chemical schemes) and realistic schemes in models was highlighted by [Pun367

and Seigneur, 2007]. One important weakness was identified to be the biogenic precursor368

emissions estimation. A first step towards improving our SOA knowledge is improvements to369

the biogenic emission inventories [Simpson et al., 2007]. Next steps concern more directly the370

representation of atmospheric processes by modeling schemes, several gaps are identified and371

may be described as follows, among others:372

1. Photochemistry of Semi-volatile organic species373

Recent studies pointed out the possible role of semi-volatile organic precursors in374

SOA formation ([Robinson et al., 2007; Donahue et al., 2006; Schauer et al., 2002]).375

These precursors evaporate during the emission dilution process and could be converted376

into the particulate phase after oxidation. The work of [Shrivastava et al., 2006] also377

suggests that aerosol emission factors could be underestimated since they are calculated378

at given temperature and dilution ratio so that a non negligible fraction could be not379

taken into account in current gas and particle inventories.380

2. Gas to particle conversion381

The exact physical and chemical pathways to secondary organic aerosol for most parent382

hydrocarbons are still uncertain and during the last ten years the condensation/sorption383

process described in [Bowman et al., 1997] has been adopted in models. However,384

nucleation burst was observed in rather clean environments such as boreal forest385

([Kavouras et al., 1998]) and in urban areas assuming co-nucleation effects with sulfuric386

acid ([Fan et al., 2006]). If in urban area absorption certainly dominates the issue is387

21

still open over forested and remote areas ([Svendby et al., 2008; Wexler and Johnston,388

2008; Bonn and Moortgat, 2002; Holmes, 2007]). [Kerminen et al., 1999] showed389

in clean environment (Arctic) a high amount of dicarboxylic acids such as glutaric,390

malonic, succinic and glutaric acid, in the coarse particles. These acids were observed391

in coarse urban and suburban aerosols corroborating the possible condensation/sorption392

of semi-volatile species onto pre-existing coarse particles in summer ([Jaffrezo et al.,393

2005]) or the possible role of in-cloud processes ([Hsieh et al., 2007; Oliveira et al.,394

2007]). Other missing processes could be mentioned such as oligomerization, oxidation395

of semi-volatile primary emissions, high/low NOx regimes.396

3. Deposition397

Particle formation by nucleation or condensation of semi- and non volatile compounds398

onto pre-existing particles involve different behaviors due to coagulation, growth and399

deposition processes. The way of modeling gas to particle conversion will have a strong400

impact on particle concentrations and compositions. Due to very high kinetic rates,401

secondary organic aerosols are rapidly produced over forested area and as reported402

by [Noll and Aluko, 2006; Petroff et al., 2008a, b] particle deposition over vegetation403

canopies is still uncertain and could be a source of model discrepancies. Moreover,404

secondary organic aerosols are partitioned between the gas phase and particles, in this405

work deposition velocities of semi-volatile organic species in the gas phase are set to406

zero. Accounting for a non zero velocity deposition for these species could largely407

affect the results.408

22

5. Conclusion409

In order to improve atmospheric composition modeling, an improved Secondary Organic410

Aerosols (SOA) scheme was implemented in the CHIMERE chemistry-transport model. This411

addition was done in parallel with the implementation of the MEGAN biogenic emissions412

inventory. Accounting for these processes allowed a clear improvement of the model results413

over the whole year of 2003 (in average and in comparison to PM10 surface data with an414

increase of 5 to 10% of temporal correlation).415

The insertion of more detailed schemes is a natural source of more model variability.416

That reveals problems not known before. In this study, a clear underestimation of OC417

concentrations was diagnosed during winter. A possible explanation is that wood burning418

emissions in Portugal, Italy, Slovakia and Hungary are missing in the model emission419

inventory as suggested by a recent modeling work ([Simpson et al., 2007]) carried out with420

different model and input data (meteorology and emissions). In addition, this work suggests421

that during the higher fire emission periods, OC concentrations from fires can be the major422

part of primary organic carbon. The contribution of SOA from fire emissions is low.423

Surprisingly, good correlations are observed between model and measurements for424

elevated sites such as Puy de Dome (France) and Schauinsland (Germany). In this specific425

case, the contribution of SOA to the total OC even in winter is quite high, showing that adding426

SOA is beneficial towards model performance.427

Finally, this work suggests that isoprene chemistry has a strong contribution to SOA428

concentrations and could explain large underestimate of OC concentrations in the southern429

23

Europe when this specific chemistry is not accounted for. The formulation of VOC oxidation430

to SOA remains too simple in the model and needs to be improved by adding aqueous and431

heterogeneous pathways, as well as taking into account the multi-step oxidation processes and432

their dependence on the NOx regime.433

Acknowledgments. This work was funded by the French Ministry in charge of Ecology434

(MEEDDAT). We thank David Simpson (EMEP, Norway) for providing us the CARBOSOL data,435

CARBOSOL was financed by European Commission (EVK2-2001-113). Yann Martinet (CITEPA,436

France) is aknowledged for his contribution in emission speciation. We also thank Alex Guenther437

(NCAR, USA) for providing us the MEGAN biogenic inventory. G. Curci has been supported by438

EU/FP6 CIRCE project.439

24

References440

Altieri, K. E., S. P. Seitzinger, A. G. Carlton, B. J. Turpin, G. C. Klein, and A. G. Marshall441

(2008), Oligomers formed through in-cloud methylglyoxal reactions: Chemical composition,442

properties, and mechanisms investigated by ultra-high resolution FT-ICR mass spectrometry,443

Atmos. Environ., 42, 1476–1490.444

Bauer, H., A. Kasper-Giebl, M. Loflund, H. Giebl, R. Hitzenberger, F. Zibuschka, and H. Puxbaum445

(2002), The contribution of bacteria and fungal spores to the organic carbon content of cloud446

water, precipitation and aerosols, Atmospheric Research, 64, 109–119.447

Bessagnet, B., A. Hodzic, R. Vautard, M. Beekmann, S. Cheinet, C. Honore, C. Liousse, and L. Rouıl448

(2004), Aerosol modeling with CHIMERE: preliminary evaluation at the continental scale,449

Atmos. Environ., 38, 2803–2817.450

Bessagnet, B., A. Hodzic, O. Blanchard, M. Lattuati, O. Le Bihan, H. Marfaing, and L. Rouıl451

(2005), Origin of particulate matter pollution episodes in wintertime over the Paris Basin,452

Atmos. Environ., 39, 6159–6174.453

Bond, T., D. Streets, K. Yarber, S. Nelson, J.-H. Woo, and Z. Klimont (2004), A technology-based454

global inventory of black and organic carbon emissions from combustion., Journal of455

Geophysical Research, 109, D14,203, doi:10.1029/2003JD003697.456

Bonn, B., and G. K. Moortgat (2002), New particle formation during α and β-pinene oxidation by457

O3, OH and NO3, and the influence of water vapour: particle size distribution studies, Atmos.458

Chem. Phys., 2, 183–196.459

Bowman, F. m., J. R. Odum, J. H. Seinfeld, and S. N. Pandis (1997), Mathematical model for460

gas-particle parttioning of secondary organic aerosols, Atmos. Environ., 31, 3921–3931.461

25

Dentener, F., et al. (2006), Emissions of primary aerosol and precursor gases in the years 2000 and462

1750 prescribed data-sets for AeroCom, Atmos. Chem. Phys., 6, 4321–4344.463

Dommen, J., et al. (2006), Laboratory observation of oligomers in the aerosol from isoprene/NOx464

photooxidation, Geophysical Research Letters, 33, doi:10.1029/2006GL026523.465

Donahue, N. M., A. L. Robinson, C. O. Stanier, and S. N. Pandis (2006), Coupled partitioning,466

dilution, and chemical aging of semivolatile organics, Environ. Sci. Technol., 40, 2635–2643,467

doi:10.1021/es052297.468

Dudhia, J. (1993), A nonhydrostatic version of the Penn State/NCAR mesoscale model: validation469

tests and simulation of an Atlantic cyclone and cold front., Monthly Weather Review, 121,470

1493–1513.471

EPA (1993), Volatile Organic Compounds (VOC)/Particulate Matter (PM) Speciation Data System472

(SPECIATE), Version 1.5, Emission Inventory Branch (MD-14), Tech. rep., Office of Air473

Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle474

Park, NC.475

Fan, J., R. Zhang, D. Collins, and G. Li (2006), Contribution of secondary condensable organics to476

new particle formation : A case study in Houston, Texas, Geophysical Research Letters, 33,477

doi:10.1029/2006GL026295.478

Gelencser, A., B. May, D. Simpson, A. Sanchez-Ochoa, A. Kasper-Giebl, H. Puxbaum, A. Caseiro,479

C. Pio, and M. Legrand (2007), Source apportionment of PM2.5 organic aerosol over480

Europe:primary/secondary, natural/anthropogenic, and fossil/biogenic origin, J. Geophys. Res.,481

112, doi:10.1029/2006JD008094.482

26

Graber, E. R., and Y. Rudich (2006), Atmospheric HULIS: How humic-like are they? A comprehensive483

and critical review, Atmos. Chem. Phys., 6, 729–753.484

Griffin, R. J., E. R. Cocker, R. C. Flagan, and J. H. Seinfeld (1999), Organic aerosol formation from the485

oxidation of biogenic hydrocarbons, J. Geophys. Res., 104, 3555–3567.486

Guenther, A., T. Karl, P. Harley, C. Wiedinmyer, P. I. Palmer, and C. Geron (2006), Estimates of global487

terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from488

Nature), Atmos. Chem. Phys., 6, 3181–3210.489

Guillaume, B., C. Liousse, R. Rosset, H. Cachier, B. Bessagnet, P. V. Velthoven, and N. Poisson (2007),490

Global modelling of internally mixed size-resolved inorganic/organic aerosols with a focus on491

carbonaceous components, Tellus B, 9, 283–302.492

Guillaume, B., C. Liousse, C. Galy-Lacaux, R. Rosset, E. Gardrat, H. Cachier, B. Bessagnet, and493

N. Poisson (2008), Modeling exceptional high concentrations of carbonaceous aerosols494

observed at Pic du Midi in spring-summer 2003:Comparison with Sonnblick and Puy de Dome,495

Atmos. Environ., in press.496

Hansen, M., R. DeFries, J. Townshend, and R. Sohlberg (2000), Global land cover classification at497

1km resolution using a decision tree classifier, International Journal of Remote Sensing, 21,498

1331–1365.499

Heinrich, J., and R. Slama (2007), Fine particles, a major threat to children, International Journal of500

Hygiene and Environmental Health, 210, 617–622.501

Henze, D. K., and J. H. Seinfeld (2006), Global secondary organic aerosol from isoprene oxidation,502

Geophysical Research Letters, 33, doi:10.1029/2006GL025976.503

Hitzenberger, R., A. Berner, H. Giebl, R. Kromp, S. Larson, A. Rouc, K. A., S. Marischka, and504

27

H. Puxbaum (1999), Contribution of carbonaceous material to cloud condensation nuclei505

concentrations in European background (Mt. Sonnblick) and urban (Vienna) aerosols, Atmos.506

Environ., 33, 2647–2659.507

Hodzic, A., R. Vautard, B. Bessagnet, M. Lattuati, and F. Moreto (2005), Long-term urban aerosol508

simulation versus routine particulate matter observations, Atmos. Environ., 39, 5851–5864.509

Hodzic, A., R. Vautard, H. Chepfer, P. Goloub, L. Menut, P. Chazette, J. Deuze, A. Apituley, and510

P. Couvert (2006), Evolution of aerosol optical thickness over Europe during the August511

2003 heat wave as seen from CHIMERE model simulations and POLDER data, Atmospheric512

Chemistry and Physics, 6, 1853–1864.513

Hodzic, A., S. Madronich, B. Bohn, S. Massie, L. Menut, and C. Wiedinmyer (2007), Wildfire514

particulate matter in Europe during summer 2003: Meso-scale modeling of smoke emissions,515

transport and radiative effects, Atmospheric Chemistry and Physics, 7, 4043–4064.516

Holmes, N. S. (2007), A review of particle formation events and growth in the atmosphere in the various517

environments and discussion of mechanistic implications, Atmos. Environ., 41, 2183–2201.518

Honore, C., et al. (2008), Predictability of European air quality: The assessment of three years519

of operational forecasts and analyses by the PREV’AIR system, J. Geophys. Res., 113,520

doi:10.1029/2007JD008761.521

Hsieh, L.-Y., S.-C. Kuo, C.-L. Chen, and Y. I. Tsai (2007), Origin of low-molecular-weight522

dicarboxylic acids and their concentration and size distribution variation in suburban aerosol,523

Atmos. Environ., 41, 6648–6661.524

Jaffe, D., I. McKendry, T. Anderson, and H. Price (2003), Six new episodes of trans-Pacific transport525

of air pollutants, Atmos. Environ., 37, 391–404.526

28

Jaffrezo, J.-L., G. Aymoz, and J. Cozic (2005), Size distribution of EC and OC in the aerosol of Alpine527

valleys during summer and winter, Atmos. Chem. Phys., 5, 2915–2925.528

Junker, C., and C. Liousse (2008), A global emission inventory of carbonaceous aerosol from historic529

records of fossil fuel and biofuel consumption for the period 1860-1997, Atmos. Chem. Phys.,530

8, 1195–1207.531

Kaiser, A., H. Scheifinger, W. Spangl, A. Weiss, S. Gilge, W. Fricke, L. Ries, D. Cemas, and532

B. Jesenovec (2007), Transport of nitrogen oxides, carbon monoxide and ozone to the533

Alpine Global Atmosphere Watch stations Jungfraujoch (Switzerland), Zugspitze and534

Hohenpeissenberg (Germany), Sonnblick (Austria) and Mt. Krvavec (Slovenia), Atmos.535

Environ., 41, 9273–9287.536

Kasper, A., and H. Puxbaum (1998), Seasonal variation of SO2, HNO3, NH3 and selected aerosol537

components at Sonnblick (3106 m a.s.l.), Atmos. Environ., 32(23), 3925–3939.538

Kavouras, I., N. Mihalopoulos, and E. Stephanou (1998), Formation of atmospheric particles from539

organic acids produced by forests, Nature, 395, 683–686.540

Kawamura, K., and O. Yasui (2005), Diurnal changes in the distribution of dicarboxylic acids,541

ketocarboxylic acids and dicarbonyls in the urban Tokyo atmosphere, Atmos. Environ., 39,542

1945–1960.543

Kerminen, V.-M., K. Teinila, R. Hillamo, and T. Makela (1999), Size-segregated chemistry of544

particulate dicarboxylic acids in the Arctic atmosphere, Atmos. Environ., 33, 2089–2100.545

Kleefeld, S., A. Hoffer, Z. Krivacsy, and S. G. Jennings (2002), Importance of organic and blackcarbon546

in atmospheric aerosols at Mace Head, on the West Coast of ireland (53’19’N, 9’54’W), Atmos.547

Environ., 36, 4479–4490.548

29

Klimont, Z., J. Cofala, I. Bertok, M. Amann, C. Heyes, and F. Gyarfas (2002), Estimating costs for549

controlling emissions of volatile organic compounds (VOC) from stationary sources in Europe.,550

Tech. rep., IIASA Interim Report IR-02-76., Laxenburg, Austria.551

Kroll, J. H., and J. H. Seinfeld (2008), Review : Chemistry of secondary organic aerosol: Formation552

and evolution of low-volatility organics in the atmosphere, Atmos. Environ., 42, 3593–3624.553

Kroll, J. H., N. L. Ng, S. M. Murphy, R. C. Flagan, and J. H. Seinfeld (2006), Secondary organic554

aerosol formation from isoprene photooxidation, Environ. Sci. Technol., 40, 1869–1877,555

doi:10.1021/es0524301.556

Langmann, B., S. Varghese, E. Marmer, E. Vignati, J. Wilson, P. Stier, and C. O’Dowd (2008), Aerosol557

distribution over Europe: a model evaluation study with detailed aerosol microphysics, Atmos.558

Chem. Phys., 8, 1591–1607.559

Lee, J.-T., J.-Y. Son, and C. Y.-S. (2007), The adverse effects of fine particle air pollution on respiratory560

function in the elderly, The Science of the Total Environment, 385, 28–36.561

Legrand, M., and H. Puxbaum (2007), Summary of the CARBOSOL project: Present and562

retrospective state of organic versus inorganic aerosol over Europe, J. Geophys. Res., 112,563

doi:10.1029/2006JD008271.564

Limbeck, A., and H. Puxbaum (1999), Organic acids in continental background aerosols, Atmos.565

Environ., 33, 1847–1852.566

Lukacs, H., A. Gelencser, H. Puxbaum, C. Pio, M. Legrand, A. Caseiro, A. Limbeck, and S. Preunkert567

(2007), Seasonal trends and possible sources of brown carbon based on two-years aerosol568

measurements at six sites in Europe, J. Geophys. Res., doi:10.1029/2006JD008151.569

Mayol-Bracero, O. L., P. Guyon, B. Graham, G. Roberts, M. O. Andreae, S. Decesari, M. C. Facchini,570

30

S. Fuzzi, and P. Artaxo (2002), Water-soluble organic compounds in biomass burning aerosols571

over amazonia: 2. apportionment of the chemical composition and importance of the polyacidic572

fraction, J. Geophys. Res., 107, doi:10.1029/2001JD000522.573

Moukhtar, S., B. Bessagnet, L. Rouıl, and V. Simon (2005), Monoterpene emissions from beech (fagus574

sylvatica) in a french forest and impact on secondary pollutants formation at regional scale,575

Atmos. Environ., 39, 3535–3547.576

Nenes, A., C. Pilinis, and S. N. Pandis (1998), ISORROPIA : a new thermodynamic equilibrium model577

for multiphase multicomponent marine aerosols, Aquatic Geochemistry, 4, 123–152.578

Noll, K. E., and O. Aluko (2006), Changes in large particle size distribution due to dry deposition579

processes, Journal of Aerosol Science, 37, 1797–1808.580

Odum, J. R., T. P. W. Jungkamp, R. J. Griffin, J. L. Forstner, R. C. Flagan, and J. H. Seinfeld (1997),581

Aromatics, reformulated gasoline and atmospheric organic aerosol formation, Environ. Sci.582

Technol., 31, 1890–1897.583

Oliveira, C., et al. (2007), Seasonal distribution of polar organic compounds in the urban atmosphere of584

two large cities from the north and south of Europe, Atmos. Environ., 41, 5555–5570.585

Petroff, A., A. Mailliat, M. Amielh, and F. Anselmet (2008a), Aerosol dry deposition on vegetative586

canopies. Part I: Review of present knowledge, Atmos. Environ., 42, 3625–3653.587

Petroff, A., A. Mailliat, M. Amielh, and F. Anselmet (2008b), Aerosol dry deposition on vegetative588

canopies. Part II: A new modelling approach and applications, Atmos. Environ., 42, 3654–3683.589

Pio, C. A., et al. (2007), Climatology of aerosol composition (organic versus inorganic) at nonurban590

sites on a west-east transect across Europe, J. Geophys. Res., 112, doi:10.1029/2006JD008038.591

Plewka, A., T. Gnauk, E. Bruggemann, and H. Herrmann (2006), Biogenic contributions to the592

31

chemical composition of airborne particles in a coniferous forest in Germany, Atmos. Environ.,593

40, S103–S115.594

Pun, B., and C. Seigneur (2007), Investigative modeling of new pathways for secondary organic aerosol595

formation, Atmos. Chem. Phys., 7, 2199–2216.596

Pun, B., C. Seigneur, and K. Lohman (2006), Modeling secondary organic aerosol via multiphase597

partitioning with molecular data, Environ. Sci. Technol., 40, 4722–4731.598

Putaud, J.-P., et al. (2004), A European aerosol phenomenology–2: chemical characteristics of599

particulate matter at kerbside, urban, rural and background sites in Europe, Atmos. Environ., 38,600

2579–2595.601

Puxbaum, H., and M. Tenze-Kunit (2003), Size distribution and seasonal variation of atmospheric602

cellulose, Atmos. Environ., 37, 3693–3699.603

Robinson, A. L., N. M. Donahue, M. K. Shrivastava, E. A. Weitkamp, A. M. Sage, A. P. Grieshop,604

T. E. Lane, J. R. Pierce, and S. N. Pandis (2007), Rethinking organic aerosols: Semivolatile605

emissions and photochemical aging, Science, 315, doi:10.1126/science.1133061.606

Schaap, M., et al. (2004), Anthropogenic black carbon and fine aerosol distribution over europe, J.607

Geophys. Res., 109, D18,207, doi:10.1029/2003JD004330.608

Schaap, M., et al. (2007), Evaluation of long-term aerosol simulations from seven air quality models609

and their ensemble in the eurodelta study, Atmos. Environ., p. submitted.610

Schauer, J. J., M. J. Kleeman, G. R. Cass, and B. R. T. Simoneit (2002), Measurement of emissions611

from air pollution sources. 5. C1-C32 organic compounds from gasoline-powered motor612

vehicles, Environ. Sci. Technol., 36, 1169–1180.613

Schlesinger, R. B., N. Kunzli, G. M. Hidy, T. Gotschi, and M. Jerrett (2006), The health relevance614

32

of ambient particulate matter characteristics: coherence of toxicological and epidemiological615

inferences, Inhal. Toxicol., 18, 96–125.616

Schmidl, C., I. L. Marr, A. Caseiro, P. Kotianova, A. Berner, H. Bauer, A. Kasper-Giebl, and617

H. Puxbaum (2008), Chemical characterisation of fine particle emissions from wood stove618

combustion of common woods growing in mid-European Alpine regions, Atmos. Environ., 42,619

126–141.620

Schmidt, H., C. Derognat, R. Vautard, and M. Beekmann (2001), A comparison of simulated and621

observed ozone mixing ratios for the summer of 1998 in western Europe, Atmos. Environ., 35,622

6277–6297.623

Schulz, M., et al. (2006), Radiative forcing by aerosols as derived from the AeroCom present-day and624

pre-industrial simulations, Atmos. Chem. Phys., 6, 5225–5246.625

Shrivastava, M. K., E. M. Lipsky, C. O. Stanier, and A. L. Robinson (2006), Modeling semivolatile626

organic aerosol mass emissions from combustion systems, Environ. Sci. Technol., 40,627

2671–2677, doi:10.1021/es0522231.628

Simpson, D., A. Guenther, C. N. Hewitt, and R. Steinbrecher (1995), Biogenic emissions in Europe. 1.629

estimates and uncertainties, J. Geophys. Res., 100D, 22,875–22,890.630

Simpson, D., K. E. Yttri, Z. Klimont, K. Kupiainen, A. Caseiro, A. Gelencser, C. Pio, H. Puxbaum, and631

M. Legrand (2007), Modeling carbonaceous aerosol over Europe: Analysis of the CARBOSOL632

and EMEP EC/OC campaigns, J. Geophys. Res., 112, doi:10.1029/2006JD008158.633

Stern, R., et al. (2008), A model inter-comparison study focussing on episodes with elevated PM10634

concentrations, Atmos. Environ., doi:10.1016/j.atmosenv.2008.01.068.635

33

Surratt, J. D., et al. (2006), Chemical composition of secondary organic aerosol formed from the636

photooxidation of isoprene, J. Phys. Chem., 110, 9665–9690.637

Svendby, T. M., M. Lazaridis, and K. Tørseth (2008), Temperature dependent secondary organic aerosol638

formation from terpenes and aromatics, J. Atmos. Chem., 59, 25–46, doi:10.1007/s10874-007-639

9093-7.640

Textor, C., et al. (2006), Analysis and quantification of the diversities of aerosol life cycles within641

AeroCom, Atmos. Chem. Phys., 6, 1777–1813.642

Tsyro, S., D. Simpson, L. Tarrason, Z. Klimont, K. Kupiainen, C. Pio, and K. E. Yttri (2007), Modeling643

of elemental carbon over Europe, J. Geophys. Res., 112, doi:10.1029/2006JD008164.644

Turpin, B. J., and H.-J. Lim (2001), Species contributions to PM2.5 mass concentrations : Revisiting645

common assumptions for estimating organic mass, Aerosol Sci. Tech., 35, 602–610.646

Van Dingenen, R., et al. (2004), A European aerosol phenomenology–1: physical characteristics of647

particulate matter at kerbside, urban, rural and background sites in Europe, Atmos. Environ., 38,648

2561–2577.649

van Donkelaar, A., R. V. Martin, R. J. Park, C. L. Heald, T.-M. Fu, H. Liao, and A. Guenther (2007),650

Model evidence for a significant source of secondary organic aerosol from isoprene, Atmos.651

Environ., 41, 1267–1274.652

Van Loon, M., et al. (2007), Evaluation of long-term ozone simulations from seven regional air quality653

models and their ensemble average, Atmos. Environ., 41, 2083–2097.654

Vautard, R., B. Bessagnet, M. Chin, and L. Menut (2005), On the contribution of natural aeolian655

sources to particulate matter concentrations in Europe: Testing hypotheses with a modelling656

approach, Atmos. Environ., 39, 3291–3303.657

34

Vautard, R., et al. (2003), Paris emission inventory diagnostics from ESQUIF airborne measurements658

and a chemistry transport model, Journal of Geophysical Research, 108(D17), 8564,659

doi:10.1029/2002JD002797.660

Vautard, R., et al. (2007), Evaluation and intercomparison of ozone and PM10 simulations by several661

chemistry-transport models over 4 european cities within the city-delta project, Atmos. Environ.,662

41, 173–188.663

Vestreng, V. (2003), Review and Revision. Emission data reported to CLRTAP, Tech. rep., EMEP664

MSC-W, http://www.emep.int/mscw/mscw publications.html.665

Wexler, A. S., and M. V. Johnston (2008), What have we learned from highly time-resolved666

measurements during EPA’s Supersites program and related studies?, Journal of the Air &667

Waste Management Association, 58, 303–319.668

Wiedinmyer, C., B. Quayle, C. Geron, A. Belote, D. McKenzie, X. Y. Zhang, S. O’Neill, and K. K.669

Wynne (2006), Estimating emissions from fires in North America for air quality modeling,670

Atmos. Environ., 40(19), 3419–3432.671

Yttri, K. E., et al. (2007), Elemental and organic carbon in PM10: a one year measurement campaign672

within the European Monitoring and Evaluation Programme EMEP, Atmos. Chem. Phys., 7,673

5711–5725.674

Zhang, Y., J.-P. Huang, D. K. Henze, and J. H. Seinfeld (2007), Role of isoprene in secondary organic675

aerosol formation on a regional scale, J. Geophys. Res., doi:10.1029/2007JD008675.676

Bertrand BESSAGNET, INERIS, Institut National de l’Environnement Industriel677

35

et des Risques, Parc technologique ALATA, 60550 Verneuil en Halatte, France.678

[bertrand.bessagnet@ineris.fr]679

Received680

This manuscript was prepared with AGU’s LATEX macros v5, with the extension package681

‘AGU++’ by P. W. Daly, version 1.6 from 1999/02/24.682

36

Figure Captions683

(a) Total terpenes (former algorithm) (b) Total terpenes (MEGAN)

(c) Isoprene (former algorithm) (d) Isoprene (MEGAN)

Figure 1. Biogenic emissions (terpenes and isoprene) for July 2003 (in Mg/cell) calculated with the

former algorithm in CHIMERE and the new MEGAN inventory.

37

Figure 2. Location of stations, red triangles : OC and EC measurements available; green circles :

PM10 measurements. For IT04, SK04, AT02 sites both PM10 and OC/EC measurements are available.

38

Figure 3. Comparisons between observed (plus symbols) and simulated (circle symbols) OC concen-

trations. Simulated SOA are represented by diamond symbols.

39

Figure 4. Average contribution (%) in 2003 of secondary organic carbon to the total organic carbon

calculated by CHIMERE.

Figure 5. Average contribution (%) in 2003 of isoprene secondary organic carbon to the total secondary

organic carbon calculated by CHIMERE.

40

Figure 6. Comparisons between observed (plus symbols) and simulated (circle symbols) EC concentra-

tions.

41

Figure 7. Impact of forest fires (in %) estimated by CHIMERE on August 1st - 15, 2003. Figures

displayed the ratio between [top] the primary OC concentrations from fires and total primary OC, and

the ratio between [bottom] SOA concentrations from fires and total SOA

42

Tables684

43

Reactions kinetic rates (molec.cm−3.s−1)685

TOL+OH→ 0.004×AnA0D + 0.001×AnA1D 1.81×10−12exp(355/T)686

+ 0.084×AnBmP + 0.013×AnBlP687

TMB+OH→ 0.002×AnA0D + 0.002× AnA1D + 0.001×AnA2D 9.80×10−9/T688

+ 0.088×AnBmP + 0.006×AnBlP689

NC4H10+OH→ 0.07×AnBmP 1.36×10−12exp(190/T)−2690

API+OH→ 0.30×BiA0D + 0.17×BiA1D + 0.10×BiA2D 1.21×10−11exp(444/T)691

API+O3→ 0.18×BiA0D + 0.16×BiA1D + 0.05×BiA2D 1.01×10−15exp (-732/T)692

API+NO3→ 0.80×BiBmP 1.19×10−12exp(490/T)693

BPI+OH→ 0.07×BiA0D + 0.08×BiA1D + 0.06×BiA2D 2.38×10−11exp(357/T)694

BPI+O3→ 0.09×BiA0D + 0.13×BiA1D + 0.04×BiA2D 1.50×10−17695

BPI+NO3→ 0.80×BiBmP 2.51×10−12696

LIM+OH→ 0.20×BiA0D + 0.25×BiA1D + 0.005×BiA2D 1.71×10−10697

LIM+O3→ 0.09×BiA0D + 0.10×BiA1D 2×10−16698

TPO+OH→ 0.70×BiA0D + 0.075×BiA1D 5.10×−8/T699

TPO+O3→ 0.50×BiA0D + 0.055×BiA1D 7.50×10−14/T700

TPO+NO3→ 0.70×BiA0D + 0.075×BiA1D 4.30×10−9/T701

ISO+OH→ 0.232×ISOPA1 + 0.0288×ISOPA2 2.55×10−11exp(410/T)702

44

Table 1. Gas phase chemical scheme for SOA formation in CHIMERE. The surrogate SOA

compounds consist of six hydrophilic species that include an anthropogenic

nondissociative species (AnA0D), an anthropogenic once-dissociative species

(AnA1D), an anthropogenic twice-dissociative species (AnA2D), a biogenic non

dissociative species (BiA0D), a biogenic once-dissociative species (BiA1D) and

a biogenic twice-dissociative species (BiA2D), three hydrophobic species that

include an anthropogenic species with moderate saturation vapor pressure (AnBmP),

an anthropogenic species with low saturation vapor pressure (AnBlP) and a biogenic

species with moderate saturation vapor pressure (BiBmP), and two surrogate compounds

for the isoprene oxidation products.

45

Station Country Lat. (oN) Long. (oE) Altitude(m)703

Puy de Dome (PDDa) France 45.45 3.00 1450705

Perigueux (PERId) France 45.19 0.73 n.a.706

K-Puszta (KPZa) Hungary 46.97 19.58 125707

Mace Head (IE31b,c) Irland 53.33 -9.90 25708

Kollumerwaard (NL09b,c) The Netherlands 53.33 6.28 0709

Payerne (CH02c) Switzerland 46.82 6.95 489710

Tanikon (CH03c) Switzerland 47.48 8.90 539711

Langenbrugge (DE02b,c) Germany 52.80 10.76 74712

Neuglobsow (DE07c) Germany 53.17 13.03 62713

Schmucke (DE08c) Germany 50.65 10.77 937714

Zingst (DE09c) Germany 54.43 12.73 1715

Schauinsland (SILa) Germany 47.92 7.90 1205716

Kosetice (CZ03b,c) The Czech Republic 49.58 15.08 534717

Stara Lesna (SK04b,c) Slovakia 49.15 20.28 808718

Liesek (SK05c) Slovakia 49.37 19.68 892719

Starina (SK06c) Slovakia 49.05 22.27 345720

Illmitz (AT02b,c) Austria 47.77 16.77 117721

St. Koloman (AT04b,c) Austria 47.65 13.20 851722

Montelibretti (IT01c) Italy 42.10 12.63 48723

Ispra (IT04b,c) Italy 45.80 8.63 209724

San Pietro Capofiume (IT08b) Italy 44.48 11.33 n.a.725

46

Station Country Lat. (oN) Long. (oE) Altitude(m)704

Braganca (PT01b,c) Portugal 41.82 -6.77 691726

Aveiro(AVEa) Portugal 40.57 -8.63 48727

Penicuick (GB46b) Great Brittain 55.95 -3.22 n.a.728

Ghent (BE02b) Belgium 51.05 3.72 n.a.729

Keldsnor (DK05b,c) Denmark 54.73 10.73 9730

Niembro (ES08b,c) Spain 43.44 -4.85 134731

Cabo de Creus (ES10b,c) Spain 42.32 3.32 23732

Barcarrota (ES11b,c) Spain 38.47 -6.92 393733

Zarra (ES12b,c) Spain 39.09 -1.10 885734

Penausende (ES13b,c) Spain 41.28 -5.87 985735

Els Torms (ES14b,c) Spain 41.40 0.72 470736

O Savinao (ES16b,c) Spain 43.23 -7.70 506737

Iskrba (SI08b,c) Slovenia 45.57 14.87 520738

Table 2. Names and coordinates of stations. (a) CARBOSOL sites (EC/OC,

diacids data), (b) EMEP sites of the EC/OC EMEP campaign, (c)

routine EMEP sites (PM10 data), (d) station from the french

monitoring network. Most of the stations are rural background sites,

except IT08, BE02 and PERI that are urban background sites.

47

Station Model Obs. RMSE Corr.

AT02a 3.86 5.82 4.16 0.107

AVEb 1.16 5.43 5.96 0.014

BE02a 2.16 3.97 2.13 0.904

CZ03a 2.84 4.98 3.59 0.328

DE02a 1.99 4.05 3.32 0.594

GB46a 0.70 1.67 1.21 0.823

IE31a 0.42 1.48 1.58 0.831

IT04a 2.49 8.30 10.22 -0.111

IT08a 2.48 5.76 4.76 -0.001

KPZb 1.33 6.46 6.38 0.254

NL09a 1.16 2.25 1.74 0.620

PDDb 0.82 1.84 1.28 0.914

PT01a 0.80 5.30 7.48 -0.081

SILb 1.74 2.47 1.34 0.750

SK04a 2.89 4.16 2.76 0.285

Table 3. Error statistics, mean model values, mean observations, Root Mean Square Errors (RMSE)

and Correlation factor for OC comparisons, (a) daily average values for the EMEP campaign and (b)

weekly average values for the CARBOSOL campaign.

48

Station Corr. without Corr. with SOA Corr. with SOA739

SOA (without isoprene SOAa) (with isoprene SOAb)740

AT02 0.626 0.623 0.600743

AT04 0.504 0.562 0.633744

CH02 0.437 0.434 0.405745

CH03 0.602 0.620 0.612746

DE07 0.803 0.806 0.808747

DE08 0.580 0.651 0.703748

DE09 0.840 0.838 0.837749

DK05 0.846 0.847 0.848750

ES08 0.583 0.664 0.704751

ES10 0.177 0.221 0.275752

ES11 0.778 0.807 0.835753

ES12 0.645 0.660 0.703754

ES13 0.730 0.754 0.781755

ES14 0.526 0.543 0.587756

ES16 0.660 0.716 0.747757

IT01 0.413 0.450 0.486758

IT04 0.607 0.572 0.506759

PERI 0.391 0.488 0.512760

SI08 0.427 0.469 0.503761

SK04 0.440 0.454 0.458762

49

Station Corr. without Corr. with SOA Corr. with SOA741

SOA (without isoprene SOAa) (with isoprene SOAb)742

SK05 0.441 0.416 0.365763

SK06 0.435 0.412 0.353764

Table 4. Impact of the SOA scheme implementation on the temporal correlation coefficients

for PM10 concentrations observed and simulated. In bold, correlation coefficients

improved by implementing the complete SOA scheme. (a) Only terpene and anthropic VOC’s

chemistry, (b) Complete scheme with isoprene chemistry.

Station Correlation

AVE 0.721

KPZ -0.136

PDD 0.646

SIL 0.512

Table 5. Temporal correlation coefficient between modeled SOA and total diacids concentrations during

the CARBOSOL campaign.

50

Station Model Obs. RMSE Corr.

AT02a 1.12 1.04 0.88 0.718

AVEb 0.23 1.09 1.03 0.303

BE02a 1.70 1.65 0.74 0.768

CZ03a 0.90 1.05 0.70 0.522

DE02a 0.73 0.55 0.82 0.598

GB46a 0.31 0.52 0.42 0.746

IE31a 0.13 0.24 0.20 0.894

IT04a 0.72 1.83 1.54 0.582

IT08a 0.58 1.29 0.94 0.467

KPZb 0.33 1.14 1.03 0.395

NL09a 0.54 0.47 0.38 0.732

PDDb 0.09 0.26 0.24 0.226

PT01a 0.11 1.03 1.44 0.295

SILb 0.34 0.30 0.27 0.171

SK04a 0.92 0.85 0.90 0.438

Table 6. Error statistics, mean model values, mean observations, Root Mean Square Errors (RMSE)

and Correlation factor for EC comparisons, (a) daily average values for the EMEP campaign and (b)

weekly average values for the CARBOSOL campaign.