CHIMERE 2013: a model for regional atmospheric …menut/pp/2013-gmdd-chimere...Menut L. et al.: CHIMERE: a model for regional atmospheric composition modelling 3 Fig. 1. General principle

Manuscript prepared for Geosci. Model Dev.- Date: 28 May 2013

CHIMERE 2013: a model for regional atmospheric compositionmodellingLaurent MENUT1, Bertrand BESSAGNET2, Dmitry KHVOROSTYANOV1, Matthias BEEKMANN3, NadegeBLOND4, Augustin COLETTE2, Isabelle COLL3, Gabriele CURCI5, Gilles FORET3, Alma HODZIC6, SylvainMAILLER1, Frederik MELEUX2, Jean-Louis MONGE1, Isabelle PISON7, Guillaume SIOUR3, Solene TURQUETY1,Myrto VALARI1, Robert VAUTARD7, and Marta G. VIVANCO8

1Laboratoire de Meteorologie Dynamique, IPSL, CNRS UMR8539 91128 Palaiseau Cedex, France2Institut national de l’environnement industriel et des risques, Parc technologique ALATA, 60550 Verneuil en Halatte, France3Laboratoire Interuniversitaire des Systemes Atmospheriques, IPSL, UMR CNRS 7583, Universite Paris Est Creteil (UPEC)and Universite Paris Diderot (UPD), Creteil, France4Laboratoire Image Ville Environnement, CNRS/UDS, UMR 7362, Strasbourg, France5Dipartimento di Scienze Fisiche e Chimiche - CETEMPS, Universita degli Studi dell’Aquila, L’Aquila, Italy6NCAR Atmospheric Chemistry Division, 3450 Mitchell Lane, Boulder, Colorado, United States7Laboratoire des Sciences du Climat et de l’Environnement, IPSL, CEA-CNRS-UVSQ, UMR 8212, Gif sur Yvette, France8Environment Department, CIEMAT, Madrid Spain.

Abstract. Tropospheric trace gas and aerosol pollutants haveadverse effects on health, environment and climate. In orderto quantify and mitigate such effects, a wide range of pro-cesses leading to the formation and transport of pollutantsmust be considered, understood and represented in numer-5

ical models. Regional scale pollution episodes result fromthe combination of several factors: high emissions (fromanthropogenic or natural sources), stagnant meteorologicalconditions, kinetics and efficiency of the chemistry and thedeposition. All these processes are highly variable in time10

and space, and their relative contribution to the pollutantsbudgets can be quantified with chemistry-transport models.The CHIMERE chemistry-transport model is dedicated toregional atmospheric pollution event studies. Since it hasnow reached a certain level a maturity, the new stable ver-15

sion, CHIMERE 2013, is described to provide a referencemodel paper. The successive developments of the model arereviewed on the basis of published investigations that are ref-erenced in order to discuss the scientific choices and to pro-vide an overview of the main results.20

1 Introduction

The rapid growth of urban areas and increased industriali-sation have created the need for air quality assessment andmotivated the first regional-scale studies on anthropogenic25

pollution in the early 1990s (Fenger (2009), among others).The first systematic measurements have been implementedby the air quality agencies in the source regions, which of-ten coincided with the most densely populated areas. One ofthe first targeted pollutants was the sulphur dioxide due to its30

effects on acid rain and forest ecosystems.

Correspondence to: Laurent Menut,[email protected]

As a result of emission reductions enforced in the indus-trial activity sector, concentrations of sulphur dioxide weregreatly reduced in the 1990s. Then the focus was shifted toother gaseous pollutants such as ozone and nitrogen oxides35

that were shown to have adverse health effects on popula-tions. More recently, particles have become a priority. Inparallel, research on biogenic pollution has long been moremodest in contrast to anthropogenic sources. Having no pos-sible influence on biogenic pollutants, the research commu-40

nity has perceived these sources, perhaps erroneously, as lessintense or less important to study.

Even if air pollution was considered as a local and mostlyurban problem, it has been shown that ozone and its precur-sors may be transported over long distances. Therefore, to45

study local pollution and represent effects of anthropogenicand biogenic emissions, chemistry and transport on the lo-cal pollution budget, models need to integrate processes overlarge spatial scales. While early models were based on statis-tical assumptions and could not account for sporadic changes50

in the atmospheric forcing, the past two decades have seenthe development of deterministic and Eulerian models. Someof these models are very complex and dedicated to field oridealised studies of a few days, over specific regions. Othersare dedicated to long-range transport only. In addition, some55

models allow fast simulations and are thus suitable for dailyforecast, Monks (2009).

The CHIMERE model has been in development for morethan fifteen years and is intended to be a modular frameworkavailable for community use. It includes the necessary state-60

of-the-art parameterisations to simulate reasonable pollutantconcentrations, but remains also computationally efficientfor forecast applications. CHIMERE is also frequently usedfor field experiment analysis studies, long-range transportand trend quantification over continental scales. Designed65

for both the research community and operational agencies,the CHIMERE model needs to be computationally stable andprovide robust results. It means that the model needs to be

2 Menut L. et al.: CHIMERE: a model for regional atmospheric composition modelling

able to estimate pollution peaks at the right time and loca-tion, but also to be able to diagnose low pollution conditions70

and avoid false alerts. As a research tool, the model needs tobe modular enough to allow adding new processes or testingspecific physico-chemical interactions. In the present paperwe describe the parameterisations included in the CHIMEREmodel and the results of recent studies explaining the ratio-75

nale of the last model improvements.This paper represents the reference model description for

the version 2013. An overview of the CHIMERE modelstructure is given in section 2. The domains definition andthe boundary conditions are described in section 3. The me-80

teorological forcings and their preprocessing are presentedin section 4 and the implementation of transport and mix-ing is discussed in section 5. The emissions taken into ac-count in the model (anthropogenic, biogenic, mineral dust,fires and the local resuspension of particulate matter) are de-85

scribed in section 6. The gaseous and aerosol chemistriesare presented in section 7. The dry and wet deposition pro-cesses are presented in section 8 and the cloud impacts insection 9. The CHIMERE model results evaluations are dis-cussed in section 10. The hybridation between model and90

observations (for sensitivity, inverse modelling and data as-similation) are presented in section 11. The experimental andoperational forecasts operated with CHIMERE are describedin section 12. Finally, a summary and new research direc-tions are presented in section 14.95

The main goal of this paper is to describe in detail all thenumerical and scientific choices and to explain the main rea-sons of these choices. In addition, ongoing and planned de-velopments and application studies are presented.

2 CHIMERE model overview100

2.1 Main characteristics

CHIMERE is an Eulerian off-line chemistry-transport model(CTM). External forcings are required to run a simulation:meteorological fields, primary pollutant emissions, chemicalboundary conditions. Using these input data, CHIMERE cal-105

culates and provides the atmospheric concentrations of tensof gas-phase and aerosol species over local to continentaldomains (from 1 km to 1 degree resolution). The key pro-cesses affecting the chemical concentrations represented inCHIMERE are: emissions, transport (advection and mixing),110

chemistry and deposition, as presented in Figure 1 . Notethat forcings have to be on the same grid and time step asthe CTM simulation. In this sense, CHIMERE is not only achemical model but a suite of numerous pre-processing pro-grams able to prepare the simulation. The model is now115

used for pollution event analysis, scenario studies, opera-tional forecast and more recently for impact studies of pol-lution on health (Valari and Menut (2010)) and vegetation(Anav et al. (2011)).

The first model version was released in 1997 and was120

a box model covering the Paris area including only gas-phase chemistry (Honore and Vautard (2000), Vautard et al.(2001)). In 1998, the model is implemented for its first fore-cast version (Pollux) during the ESQUIF experiment (Menutet al. (2000b)) still over the Paris area (Vautard et al. (2000)).125

At the same time, the adjoint model was developed to es-timate the sensitivity of concentrations to all parameters(Menut et al. (2000a)). In 2001, the geographical domain wasextended over Europe with a cartesian mesh (Schmidt et al.(2001)) and the new experimental forecast platform (PIO-130

NEER) is set up. In 2003, the experimental forecast becameoperational with the PREVAIR system operated at INERIS(Honore et al. (2008), Rouıl et al. (2009)). The aerosol mod-ule is implemented in 2004 (Bessagnet et al. (2004)) with fur-ther improvements concerning the dust natural emissions and135

resuspension over Europe (Vautard et al. (2005), Hodzic et al.(2006a) Bessagnet et al. (2008)) and evaluated against long-term and field measurements (Hodzic et al. (2005), Hodzicet al. (2006b)). The development of the mineral dust ver-sion started in 2005 (Menut et al. (2005b)). Chemistry was140

not included in that version and a new horizontal domain hadbeen designed to cover the whole northern Atlantic and Eu-rope, including the Saharan desert and downwind regions. In2006, an important step is achieved with the development ofthe parallel version of the model and its first implementation145

on a massively parallel computer (the ECMWF computer inthe framework of the FP6/GEMS project).

The CHIMERE model is now considered as a state-of-the-art model. It has been involved in numerous intercompar-ison studies mainly focusing on ozone and PM10 from the150

urban scale (Vautard et al. (2007); Van Loon et al. (2007);Schaap et al. (2007)) to continental scale (Solazzo et al.(2012b), Zyryanov et al. (2012)). Moreover, the model hasbeen mainly applied over Europe, but also more recently overAfrica and the North Atlantic for dust simulations, over Cen-155

tral America during the MILAGRO project to study organicaerosols (Hodzic et al. (2009), Hodzic et al. (2010b), Hodzicet al. (2010a)) and over the US within the AQMEII project(Solazzo et al. (2012b)).

Finally, the development of CHIMERE follows three main160

rules. First, concentrations of main pollutants are calcu-lated with the best possible accuracy using well evaluatedand state-of-the-art parameterisations. Second, a modularframework is maintained to allow updates to the code by de-velopers but also all interested users. Third, the code is kept165

computationally efficient to allow long-term simulations, cli-matological studies and operational forecast.

2.2 The CHIMERE software

In order to facilitate software distribution, CHIMERE is pro-tected under the General Public License. This paper presents170

the last model version release called CHIMERE 2013. Thesource code and the associated documentation is available

Menut L. et al.: CHIMERE: a model for regional atmospheric composition modelling 3

Fig. 1. General principle of a chemistry-transport model such as CHIMERE. In the box ’Meteorology’, u∗ stands for the friction velocity,Q0 the surface sensible heat flux, L the Monin-Obukhov length and BLH the boundary layer height. cmod and cobs are for the chemicalconcentrations fields for the model and the observations, respectively.

on a web site www.lmd.polytechnique.fr/chimere. The docu-mentation is both technical and scientific. It includes a chap-ter dedicated to the set-up of a test case simulation that allows175

new users to easily carry out a CHIMERE simulation: modelconfiguration and data (meteorology and emissions files) areprovided to simulate the 2003 heat wave in western Europe.

CHIMERE is a National Tool of the French Insti-tut des Sciences de l’Univers, meaning that support has180

to be provided to the users of model. Two mailinglists exist for this support: [email protected] send questions to the model developers and [email protected] to initiate discussions, ex-change programs or data between users. In addition, two-185

day training courses are organised twice a year. Each trainingcourse is free of charge for participants and offers a completetraining to be able to install the code, launch a simulation andchange surface emissions or other parameters in the code.

The code is completely written in Fortran90, and running190

scripts are written in shell (using gnu-awk for input datafilesprocessing). The required software is a Fortran 95 compiler(g95 and gfortran are both free and efficient, but Makefileswith Intel’s ifort compiler options are also provided). The re-quired libraries are NetCDF (either 3.6.x or 4.x.x), MPI (see195

below), and GRIB API (associated to the use of the ECMWFmeteorological datasets). The model includes tools that canhelp the user to configure the model’s Makefiles for the li-braries already installed.

The model computation time for one AMDx64 node of200

16 CPUs is 1h30min for 1 month of simulation for the Paris

area, at 15 km resolution, the domain size being 45x48x8with a time step of 360 seconds on average. CHIMERE usesthe distributed memory scheme, and MPI message passinglibrary. It is maintened for Open MPI (recommended) and205

LAM/MPI, but works, with minor changes in the scripts,with MPICH or other MPI compatible parallel environ-ments. The model parallelism results from a Cartesian di-vision of the main geographical domain into several sub-domains, each one being processed by a worker process.210

Each worker performs the model integration in its geographi-cal sub-domain as well as boundary condition exhanges withits neighbours. In addition a master process performs initial-isations and file input/output. To configure the parallel sub-domains, the user has to specify two parameters in the model215

parameter file: the number of sub-domains for the zonal andmeridional directions. The total number of CPUs used istherefore the product of these two numbers, plus one CPUfor the master process.

For graphical postprocessing, simple interfaces are avail-220

able using either the GMT 5 or the GrADS 6 free software.Also an additional graphical user interface (GUI) softwareCHIMPLOT is provided. It allows making various 1D or 2Dplots (e.g., longitude-latitude or time-altitude maps, verticalslices, time series, vertical profiles). One can also overlay225

multiple fields (e.g., O3 concentrations, wind vectors, andpressure contours) and perform simple operations such ascalculating daily maxima, daily means, vertical or horizon-tal averaging or integrations.


2.3 The CHIMERE code organisation230

Fig. 2. General CHIMERE structure for time integration of all pro-cesses

Apart from the pre-processing, i.e., the preparation ofthe forcing, the CHIMERE model is split into two mainparts: an initialisation phase and the model integration phase( Figure 2 ).

The initialisation phase consists in reading all input param-235

eters as well as the preparation of the initial meteorologicaland chemical field.

The model integration phase is split into three stages:

– A hourly time-step that corresponds to the provision offorcings,i.e meteorological fields, emission fluxes and240

chemical boundary conditions.

– A user’s defined coarse time step ’nphour’, correspond-ing to the time interpolation of ”physical” parameters,such as wind, temperature, reactions rates etc. In paral-lel, to optimise the time simulation and prevent issues245

associated to the violation of the Courant-Friedrich-Levy (CFL) criteria, a ”physical” time step is dynami-cally estimated in the meteorological pre-processor tak-ing into account horizontal and vertical winds and deepconvective updraft, as presented in Figure 3 . During250

the run, if the specified time step is longer than the rec-ommended time step, the recommended time step willbe used in model integrations. If the user’s definedtime-step is shorter as the recommended one, the user’schoice is applied (even if this is not the optimal choice).255

– A user’s defined fine time step ’ichemstep’: this corre-sponds to the integration of the chemical mechanism,

including concentration’s increments due to all pro-cesses. This is achieved by the two-step scheme. Due tothe stiffness of the chemical system to solve, this time260

step must be at least 30s (or less if possible). In prac-tice, for large domains, such as Western Europe, a ”veryquick formulation” with a 10-minute physical step andno sub-chemical steps, i.e. all processes are stepped to10 minutes, is realistic. It is possible to select one or two265

Gauss-Seidel iterations, but the use of two iterations isstrongly recommended even if it increases the computertime by two.

Fig. 3. Calculation of the number of integration steps per hour torespect the Courant-Friedrich-Levy (CFL) number over a complete120 hours simulation. The CFL is estimated using three parame-ters: the mean wind speed |U |, the vertical wind speed in the envi-ronment w and the vertical velocity in the updraft in case of activedeep convection.

The numerical integration of all processes follows aproduction-loss budget approach as presented in Figure 4 .270

This means that all production and loss terms, for eachchemical species are calculated simultaneously to avoid errorpropagation generally created with operator-splitting tech-niques (the concentration evolution being dependent on theseveral terms order, McRae et al. (1982)). Further advan-275

tages of the scheme are 1) its stability even for quite longtime steps due to the implicitness of the formulation and 2)the simplicity of the code which facilitates the developmentof secondary models (adjoint, tangent linear).

The numerical method used to estimate the temporal so-280

lution of the stiff system of partial differential equations isadapted from the second-order TWOSTEP algorithm orig-inally proposed by Verwer (1994) for gas phase chemistryonly. It is based on the application of a Gauss-Seidel iter-ation scheme to the 2-step implicit backward differentiation285

(BDF2) formula:

cn+1 =43cn− 1

3cn−1 +

23

∆tR(cn+1) (1)


with cn being the vector of chemical concentrations attime tn, ∆t the time step leading from time tn to tn+1 andR(c) = c=P (c)−L(c)c the temporal evolution of the con-290

centrations due to chemical production and emissions (P )and chemical loss and deposition (L). Note that L is a di-agonal matrix here. After rearranging and introducing theproduction and loss terms this equation reads:

cn+1 =(I+

23

∆tL(cn+1))−1

295

×(

43cn− 1

3cn−1 +

23

∆tP (cn+1))

(2)

The implicit nonlinear system obtained can be solved per-tinently with a Gauss-Seidel method (Verwer (1994)).

Fig. 4. Principle of ’operator-splitting’ versus Chimere integration

In CHIMERE the production and loss terms P and L inequation 2 are replaced by the modified terms P = P +300

Ph+Pv and L=L+Lh+Lv , respectively. Ph and Pv de-note the temporal evolution of the concentrations due to hori-zontal (only advection) and vertical (advection and diffusion)inflow into a given grid box, Lh and Lv gives the temporalevolution due to the respective outflow divided by the con-305

centration itself.

3 Domains and boundary conditions

3.1 Domains geometry

Model domains are defined by their grid cell centres. Theuser controls the model grid through a simple longitude / lat-310

itude ASCII file, in decimal degrees. The input meteorolog-ical fields are automatically interpolated on the CHIMEREgrid. Each pair of coordinates stands for a grid cell centre,described (from the top to the bottom of the file) from Westto East then from South to North.315

In the definition of a new CHIMERE domain, theuser must check carefully whether the domain is quasi-rectangular. Most projections work, including a regular gridin geographic coordinates (longitude-latitude), provided theresolution is not too coarse (more than ≈ 2 degrees). The320

model grid can be any quasi-rectangular grid with a slowlyvarying spatial step. In case of use of the Parabolic PiecewiseMethod for horizontal transport, the grid size being consid-ered as constant in each direction locally (over 5 consecutivecells), it is recommended to define a rectangular grid. The325

sphericity effects, although taken into account, are thereforelinearized.

The model uses any number of vertical layers described inhybrid sigma-p coordinates. The pressure in hPa at the top ofeach layer k is given by the following formula:330

Pk = ak105 +bkPsurf (3)

where Psurf is the surface pressure and the hybrid coef-ficients ak and bk must be provided by the user in an inputfile. The vertical grid can be defined automatically given thepressure at the top of the model and sigma at the top of the335

surface layer. In this case, a pre-processor available in themodel is used and the resolution varies upward in a geomet-ric progression.

Table 1 gives examples of domains on which CHIMEREhas been run and/or evaluated. These domains range from340

hemispheric scale for the simulation of long-range dust trans-port with a horizontal resolution of 50km to simulations atthe scale of the urban area (Valari et al. (2011), Menut et al.(2013)) with a resolution of about 5km. The most frequentlyused vertical discretisation consists of 8 vertical levels, from345

sigma-level 0.995 to the 500hPa pressure level, indepen-dantly of the horizontal resolution, which is the configura-tion of the PREVAIR operational forecast system Rouıl et al.(2009). For other applications, such as long-range transportof dust or volcano ashes, which require a better representa-350

tion of the free troposphere, more levels are used (15 levelsup to 200hPa in Menut et al. (2009a), 18 levels up to 200hPain Boichu et al. (2013)). However, regarding the modellingof anthropogenic pollution, it has been shown that increasingthe number of levels from 8 to 20 levels does not modify sub-355

stantially the modelled concentration at ground levels. Thethickness of the first model layer is however a critical param-eter, and adding a new vertical layer close to the surface at


the sigma-level 0.999 (instead of 0.995) significantly impactsthe modelled concentrations, particularly during nightime in360

the case of very stable boundary layers Menut et al. (2013).

3.2 Landuse types

Land use types, or categories, are needed by CHIMERE tocalculate a number of processes, such as deposition, bio-genic emissions, or surface layer momentum and heat trans-365

fer. Land use files need to be constructed only once permodel domain. There are currently 9 land use categories inCHIMERE. Those categories are calculated from availableglobal land use databases, which can contain different num-ber of classes. The source code version comes with land use370

data and interfaces for two databases: the Global Land CoverFacility (GLCF) and the GlobCover Land Cover (LC).

GLCF is a 1km×1km resolution database from the Uni-versity of Maryland, following the methodology of Hansenand Reed (2000). This global land cover classification is375

based on the imagery from the AVHRR satellites analysedto distinguish 14 land cover classes. The GlobCover LC isa global land cover map at 10 arc second (300 meter) res-olution (Bicheron et al. (2011)). It contains 22 global landcover classes defined within the UN Land Cover Classifica-380

tion System (LCCS). GlobCover database is based on theENVISAT satellite mission’s MERIS sensor (Medium Res-olution Image Spectrometer) Level 1B data acquired in FullResolution (FR) mode with a spatial resolution of 300 me-ters. GlobCover LC was derived from an automatic and385

regionally-tuned classification of a time series of MERIS FRcomposites covering the period December 2004-June 2006.The global land cover NetCDF files are provided along withthe CHIMERE distribution.

The nine CHIMERE land use types are described in390

Table 2 . The correspondance table between the databaselanduse types and the CHIMERE landuse types is providedwith the model. The user can choose either GLCF or Glob-Cover by simply selecting a flag; a dedicated sequenceof scripts and programs prepares the landuse file in the395

CHIMERE format. An additional class ”inland water” hasbeen added to the classifications in both land cover databasesto distinguish between the sea water and fresh water. Thisfeature was required to avoid model emissions of the seasalt over the fresh water surfaces. The distinction was per-400

formed using a land-sea mask. So instead of the original14 GLCF and 22 GlobCover classes, the CHIMERE lan-duse pre-processor relies on 15 and 23 classes for GLCF andGlobCover, respectively. An example of CHIMERE regrid-ded landuse is displayed in Figure 5 .405

3.3 Boundary and initial conditions

As a limited area model, CHIMERE requires chemical ini-tial and boundary conditions. The boundary conditions arethree-dimensional fields covering the whole simulation pe-

Fig. 5. Example of GLCF landuse regridded over a CHIMERE do-main: the western European domain used for the GEMS projectwith a horizontal resolution of 0.5 × 0.5 degrees. For each cell thedominant landuse is shown. The color code correspond to the lan-duse number (with 1 between 0.5 and 1.5, for example). The codesare: 1. Agricultural land / crops, 2. grassland, 3. barren land, 4.inland water, 5. urban, 6 shrubs, 7 needleleaf forest, 8 broadleafforest, 9 ocean.

riod. These fields provide the concentrations of chemical410

species (gaseous and particulate) at the lateral and upper lay-ers of the CHIMERE simulation domain. Some CTMs usetabulated vertical profiles derived from observational clima-tologies. Considering that observation-based boundary con-ditions are too restrictive in terms of available species, as415

temporal and spatial coverage, CHIMERE get its boundaryconditions from global CTMs. The sensitivity study Szopaet al. (2009) illustrates the importance of using a domain thatis large enough to minimize boundary effects and allow forrecirculations within the CHIMERE domain.420

To ensure the best possible simulation quality, a commonpractice is to use the nesting option of CHIMERE, witha coarse domain to provide the most consistent boundaryconditions for the smaller nested domains. An example ofsuch a configuration is displayed in Figure 6 for a spe-425

cific study in the Paris area with the horizontal resolutiondx= 5 km. The local simulation is nested into a larger do-main with dx= 15 km, which itself is nested into a domainwith dx= 45 km. Only the largest domain makes use of ex-ternal boundary conditions ( Figure 6 ).430

Szopa et al. (2009) also investigated the sensitivity of themodel to the temporal increment of the boundary conditions.They found that using a time-variable large scale forcing im-proves the variability at the boundary of the domain com-pared to the monthly average, but the magnitude of the sen-435

sitivity decreases towards the centre of the domain. Schereet al. (2012) confirmed this finding during the AQMEII in-


Domain description Purpose Domain scale ∆x×∆y nv ptop ReferenceNorth-Atlantic Dust transport Hemispheric 1◦×1◦ 15 200 Menut et al. (2009a)Europe Operational forecast Continental 0.5◦×0.5◦ 8 500 Rouıl et al. (2009)Europe Models comparisons Continental 25km×25km 9 200 Solazzo et al. (2012b)North-America Models comparisons Continental 36km×36km 9 200 Solazzo et al. (2012b)France Operational forecast National 0.15◦×0.1◦ 8 500 Rouıl et al. (2009)Western Europe Sensitivity study Continental 45km×45km 8, 9, 20 500 Menut et al. (2013)Northern France Sensitivity study Regional 15km×15km 8, 9, 20 500 Menut et al. (2013)Ile-de-France region Sensitivity study Urban area 5km×5km 8 500 Valari et al. (2011)

Table 1. Examples of some domains for which CHIMERE simulations have been performed, with their horizontal resolution, ∆x and ∆y,their number of vertical levels, nv and the domain top, ptop, in hPa.

# Description # Description1 Agricultural land / crops 6 Shrubs2 Grassland 7 Needleleaf forest3 Barren land/bare ground 8 Broadleaf forest4 Inland Water 9 Ocean5 Urban

Table 2. Landuse categories used in CHIMERE

Fig. 6. Example of simulation domains for CHIMERE, and corre-sponding surface ozone concentrations maps. The largest domain(dx=45km) uses a global model climatology for boundary condi-tions, and forces itself the medium domain, dx=15km over the Northof France, itself forcing the small domain, dx=5km, over the Parisarea.

tercomparison exercise. Colette et al. (2011) argued that theselection of the large scale model had a larger impact than itstemporal resolution.440

Depending on the research projects that have been con-ducted in the past, preprocessing tools have been imple-mented to build boundary conditions from a variety ofglobal models. The most widely employed is LMDz-INCA(Folberth et al. (2006)), for which a climatology (average445

monthly fields) is available on the CHIMERE website. Al-

ternatively, the MOZART model was also used, for instancefor the GEMS (Hollingsworth et al. (2008)), MACC II, andAQMEII (Rao et al. (2011)) projects, and an interface withOsloCTM2 (Sovde et al. (2008)) was also developed for the450

CityZen project. For the specific case of mineral dust, theGOCART (Ginoux et al. (2001)) monthly average fields arealso available. The current version of the model includes inits namelist a series of flags in order to define which modelis to be used for the following three groups of species: gases,455

dust aerosols, and non-dust aerosols. For each of them eitherclimatological or time-varying fields can be used.

The initial conditions are a three-dimensional field corre-sponding to the starting date of the simulation. The sameglobal model fields used for boundary conditions can be used460

to initiate the simulation. If there are no global model fieldsavailable, it is also possible to start a simulation with zeroconcentrations for all species. In this case, the spin-up timeof the simulation has to be adjusted to the domain size: froma few days for a local domain to one month for a continen-465

tal domain (to take into account the long-range transport andpossible recirculations).

4 Meteorology

CHIMERE is an off-line chemistry-transport model drivenby meteorological fields, e.g., from a weather forecast model,470

such as WRF or MM5. CHIMERE contains a meteorologi-cal pre-processor that prepares standard meteorological vari-ables to be read by the model core.

The input meteorological data are processed in two stages,as presented in Figure 7 . This choice constitutes a strength475

of the CHIMERE model: the user can use CHIMERE pro-viding only very basic meteorological variables: wind speed,temperature, humidity and pressure. In this case, a completesuite of diagnostic tools for all other mandatory variables(turbulent fluctuations and fluxes) may be used. On the other480

hand, if the meteorological model provides all the necessarymeteorological variables, the meteorological interface onlyinterpolates data on the CHIMERE grid, with an hourly timestep. The diagnostic interface may also be used even if tur-bulent parameters are provided with the meteo model. The485


Fig. 7. Treatment of meteorological fields: two options are avail-able (i) using a meteorological dataset restricted to the mean pa-rameters (u,v,T,q,P, precipitation) and (ii) using a complete meteo-rological dataset that includes turbulent parameters.

user can decide to bypass turbulent parameterisations of theinput meteorological model and make use of the CHIMEREdiagnostics in order to increase the consistency of the forcingfields accross a range of input models.

The meteorological interface first transforms original vari-490

ables from any input spatial grid and temporal frequency ontostandard variables given on the CHIMERE horizontal grid ashourly values. The operations performed include: horizontaland temporal interpolation, wind vector rotation, temporaldeaccumulation of precipitations, transformation from per-495

turbation and mean values to full values, etc. The vertical in-terpolation is performed at a later stage, since a higher verti-cal resolution might be required for the turbulence and fluxesdiagnostics. In the current CHIMERE version, meteorolog-ical interfaces are provided for ECMWF (ERA-INTERIM,500

IFS), WRF (Skamarock et al. (2007)) and MM5 models.If not all required fields are provided, the pre-processor di-

agnostic model is used. It takes meteorological variables andtransforms them into variables necessary for the CHIMEREcore. These parameters are (i) the radiation attenuation, (ii)505

the boundary layer height, (iii) the friction velocity u∗, theaerodynamical resistance ra, the sensible heat flux Q0, theMonin-Obukhov length L and the convective velocity w∗.

For the photochemistry, the cloud liquid water content isnecessary to estimate the radiative attenuation. If rain water510

or ice are available, they are added to the cloud water for theattenuation effects. Note that for gaseous species (such asHNO3) and aerosols calculations, additional parameters re-lated to convective and large-scale precipitation are requiredfor the scavenging.515

Finally, and since most large-scale weather models do notinclude any ”urban parameterisation”, the possibility of cor-recting the wind speed in the surface layer (due to increasedroughness) in urban areas is offered. This will automaticallybe balanced by a vertical wind component calculated in the520

mass balance (see vertical transport below). This correctionhas however no effect at continental scale where the frac-tion of urban areas in the model grid cells are limited (seeFigure 5 for example, where the Paris city, a urban site, is

not a ’dominant’ landuse for the corresponding cell). This ur-525

ban parameterisation has however a strong impact on urbanversions of the model, mostly for primary pollutants.

4.1 Diagnostic of turbulent parameters

The turbulent parameters may be read form the meteoro-logical driver or diagnosed in CHIMERE, as explained in530

Figure 7 . In case of a CHIMERE diagnostic, the followingvariables are calculated: the friction velocity u∗, the surfacesensible heat fluxQ0, the vertical convective velocity w∗, theboundary layer height h, the bulk Richardson number RiB ,the Monin-Obukhov length L and the vertical diffusivity pro-535

file Kz .The friction velocity u∗ is used for the deposition and

calculation of diffusivities. It is a particularly sensitive pa-rameter for ozone in summer through the calculation of theaerodynamic resistance ra. Friction velocity is thus sensi-540

tive to the land use type, which is critical to deposition. Inlarge scale meteorological models, roughness lengths are of-ten too coarse for the implementation of high-resolution de-position. Therefore an update of u∗ is proposed following theLouis et al. (1982) formulation which is particularly robust.545

We recommend to use this alternative formulation, no matterwhether u∗ is available in the input fields, in order to havea deposition that is consistent with the high-resolution landuse.

u∗=√C2DN Fm |U |2 (4)550

with Fm is the Louis et al. (1982) stability function andCDNthe momentum transfer coefficient as:

CDN =k

ln

(z

z0m

) (5)

with k= 0.41 the Karman constant, z the altitude wherethe wind speed |U | is known and z0m, the momentum rough-555

ness length. The momentum stability function Fm is esti-mated depending on the bulk Richardson number,Rib, value.Rib is estimated at each altitude z as:

Rib(z) =g z

θv(z)θv(z)−θv(z0)|U |(z)2

(6)

with g= 9.81m2/s2 the gravitational acceleration and θv the560

virtual potential temperature in Kelvin.


• In unstable cases, if Rib< 0:

Fm = 1− 2b Rib

1+3b c C2DN .

√z

z0m

√|Rib|

(7)

• In stable cases, if Rib> 0:

Fm =1

1+2b Rib√1+d Rib

(8)565

with the constant b= c= d= 5. In neutral case, i.e Rib =0, Fm=1.

The surface sensible heat flux, Q0, is used to compute w∗and therefore mixing, and the height of the boundary layer.In fact only the virtual heat flux is required, which can be re-570

computed from an empirical formula Priestley (1949) usingtemperatures in the first model layers. This formula is notvery accurate. It is strongly advised to use heat fluxes fromthe meteorological model if available. If the surface sensibleheat flux Q0 is provided by the meteorological model, it is575

directly used for the computation of the convective velocityw∗:

w∗=(g Q0 h

ρ Cp θv

)1/3

(9)

where h is the convective boundary layer height, Cp thespecific heat of air at constant pressure, θv the mean virtual580

potential temperature in the first model vertical level.The Monin-Obhukov length is estimated as:

L=−θv u3

∗k g Q0

(10)

The boundary layer height (h) is derived from differentformulation depending on the atmospheric static stability.585

When stable, i.e when L> 0, h is estimated as the altitudewhen the Richardson number reaches a critical number herechosen as Ric = 0.5, following Troen and Mahrt (1986).

In unstable situations (i.e. convective), h is estimated us-ing a convectively-based boundary layer height calculation.590

This is based on a simplified and diagnostic version of theapproach of Cheinet and Teixeira (2003) which consists inthe resolution of the (dry) thermal plume equation with dif-fusion. The in-plume vertical velocity and buoyancy equa-tions are solved and the boundary layer top is taken as the595

height where calculated vertical velocity vanishes. Thermalsare initiated with a non-zero vertical velocity and potentialtemperature departure, depending on the turbulence similar-ity parameters in the surface layer.

Once the depth of the boundary layer is computed, verti-600

cal turbulent mixing can be applied following the K-diffusionframework following the parameterisation of Troen andMahrt (1986), without counter-gradient term. In each model

column, the diffusivity coefficient profile Kz (m2/s) is calcu-lated as:605

Kz = kwsz

h

(1− z

h

)2

(11)

where ws is a vertical scale given by similar formulas:

• In the stable case (when the surface sensible heat flux isnegative):

ws =u∗

(1+4.7z/L)(12)610

• In the unstable case:

ws = (u3∗+2.8ew3

∗)1/3 (13)

where e=max(0.1,z/h). A minimalKz is assumed, witha value of 0.01 m2/s in the dry boundary layer and of 1 m2/sin the cloudy boundary layer. Kz is capped to a maximal615

value ofKz=500 m2/s to avoid unrealistic mixing. Above theboundary layer, a fixed value of Kz=0.1 m2/s is prescribed.As for many CTMs considered as numerically diffusive, hor-izontal turbulent fluxes are not considered.

4.2 Diagnostic of deep convection fluxes620

Deep convection occurs when cumulus or cumulonimbusclouds (referred to as convective clouds) are present. Theseclouds are formed when air masses are unstable, when warmair is at the surface or cold air is transported in upper layers(cold front). High vertical wind speed are observed leading625

to large cloud structures along the vertical direction. On theother hand, when clouds are only due to mechanical forcings(mountains, warm fronts), they are referred to as stratiformclouds and generally exhibit low vertical velocity.

Air masses shall be quickly mixed in the troposphere when630

convective unstabilities occur under a cloud. To describe thisphenomenon, mixing schemes generally consider a cloud(and the whole column including this cloud) and its environ-ment. In the main part of deep convection parameterisations,the hypothesis of a small cloud surface compared to the total635

studied surface is used.Under the cloud, updrafts and downdrafts are observed.

The updraft originates from air masses lighter than their en-vironment when downdrafts represent the downfall of colderair (often with rain). In the updraft and the downdraft, air640

may be exchanged between the cloud and its environment.Entrainment refers to the air that flows from the environmentinto the cloud, detrainment refers to the air that flows fromthe cloud towards the environment. In order to ensure massconservation, a compensatory subsidence is observed in the645

environment.In CHIMERE, the hourly fluxes of entrainment and de-

trainement in the updrafts and the downdrafts are estimatedduring the meteorological diagnostic stage using the Tiedtkescheme, Tiedtke (1989). This scheme has been implemented650


in order to compute convective mass fluxes if any convec-tive fluxes are available from the meteorological model cou-pled with CHIMERE. The work has been done following thestudy of Olivie et al. (2004) who showed that the fluxes com-puted off-line with this methodology in TM3 give similar655

fluxes than the archived ECMWF ERA40 convective massfluxes.

The convective mass fluxes are given by:

M(z) = ρ(z)(aup wup(z)+adw wdw(z))

= −ρ(z) aenvwenv(z) (14)660

with wup(z), wdw(z) and wenv(z) the vertical wind speedin the updrafts, downdrafts and environment, respectively.aup and aenv are the fractions of coverage area. The verticalgradient of this mass flux is:

∂M(z)∂z

(z) =E(z)−D(z) (15)665

with E(z) and D(z) are the rates of mass entrained anddetrained per unit length (kg m−2 s−1). This equation worksboth for updrafts and downdrafts, with, for example for theupdraft:

E(z) = Eup(z)+Edw(z)670

D(z) = Dup(z)+Ddw(z) (16)

Fig. 8. Entrainment and detrainment fluxes in the updrafts anddowndrafts.

Figure 8 shows an example of vertical profile of theconvective fluxes; i.e. updraft and downdraft mass fluxes,and entrainment and detrainment fluxes due to the updraftand downdraft mass fluxes. In order to use this convection675

scheme, new calculations were added to the meteorologi-cal pre-processor, providing an estimate of the vertical windspeed (independently of the input from the meteorologicalmodel). In the CHIMERE model itself, these fluxes are usedto estimate mass fluxes for the pollutant species as follows:680

∂Mup(z)cup(z)∂z

=Eup(z)cenv−Dup(z)cup

∂Mdw(z)cdw(z)∂z

=Edw(z)cenv−Ddw(z)cdw (17)

with the concentrations cup, cdw and cenv in the updraft,downdraft and environment, respectively. The calculation isdone from surface to the top of the domain in order to ensure685

mass conservation.

5 Transport and mixing

For a given chemical species with concentration c, the fol-lowing conservation equation is numerically solved:

∂t(cρ)+∂iFi = 0 (18)690

with ρ is the air density and F indicating the mass fluxcorresponding to velocity v:

F = ρcv (19)

As CHIMERE is designed for structured grids where thegrid cells are nearly parallelepipedic, this equation can be695

discretized and solved separately for each of the three or-thogonal directions: zonal, meridional and vertical. Thisstrategy is known as operator splitting. Even though the useof operator splitting may generate some numerical problems,this technique is very widely used in weather modelling and700

chemistry-transport modelling, both because it is more com-putationnally efficient and also sometimes more stable andaccurate than a bi- or tri-dimensional approach, particularlywhen it is applied to high-order models (Byun et al. (1999)).Therefore, the tendency ct in the concentration c is equal to705

c(1)t +c

(2)t +c

(3)t , where c(i)t is the time derivative of concen-

tration due to transport in the ith direction. It is also assumedthat the time variations of ρ are much slower than the timevariations of c (|cρt|� |ρct|), so that Eq. 18 becomes:

ρc(i)t =−∂iFi (20)710

After time and space discretisation, if we note δ(i)c thevariation of c due to transport in direction i, the discretizedtransport equations are the following:

δ(i)c=−

(F in+ 1

2(t)−F i

n− 12(t)

∆x

)∆t (21)

The concentration increments are calculated successively715

for each direction from the initial concentration field (parallel


strategy). This equation secures mass conservation becausethe inward and outward fluxes cancel out in each direction.The time integration is of order 1. The key issue in solvingthis equation is the estimation of the fluxes at the cell inter-720

faces (F in± 1

2(t)). The way these fluxes are estimated numer-

ically determines the characteristics of the transport scheme(scheme type and order, diffusivity, numerical stability etc.).

In chemistry-transport models, various types of numeri-cal schemes have been tested and are proposed to users to725

solve the advection equations. These numerical schemesrange from simple order-1 numerical schemes such as theclassical upwind method to higher-order methods such asthe Piecewise-parabolic method (Colella and Woodward(1984)) proposed in CHIMERE and CMAQ, the Yamartino-730

Blackman cubic scheme, proposed in CMAQ (Byun et al.(1999)), or the Walcek scheme in BRAMS (Freitas et al.(2011)).

5.1 Horizontal transport

In the horizontal directions, it is assumed that the grid cell735

length does not vary substantially from one grid cell to itsneighbours. As in the CHIMERE model, species concen-trations and meteorological variables are represented on thesame grid, the wind speed at the interfaces is interpolated lin-early from the wind speeds at the centers of the two gridcells740

separated by the interface. The wind speed at the interfacewill be noted un± 1

2(t), and is assumed constant during each

coarse time step. In the following we will drop the super-script i to indicate the direction. The definition of the windspeed at the cell interface is independent of the scheme used,745

so that the definition of the fluxes at the interfaces only de-pends on the evaluation of the concentration at the interface.

Three horizontal transport schemes are available in themodel.

5.1.1 Upwind scheme750

In the upwind scheme (Courant et al. (1952)), the fluxes atthe cell interfaces are defined by the following equations ac-cording to the sign of the wind speed at the cell interface:

• If un+ 12(t)> 0

755

Fn± 12(t) = cnρnun+ 1

2(t) (22)

• If un+ 12(t)< 0

Fn± 12(t) = cn+1ρnun+ 1

2(t) (23)

The assumption made in this scheme is that the tracer con-760

centration is uniform in each grid cell, so that the mass fluxat the interface is the product of the wind at the interface bythe tracer concentration in the upwind cell.

5.1.2 Van Leer scheme

The Van Leer scheme used in CHIMERE is commonly called765

Van Leer I because it is the first one described in the semi-nal paper of Van Leer (1979). This scheme is of order 2 inspace, it assumes that the concentration inside a grid cell isdescribed by a linear slope between the two cell interfaces:

cn(x) = cn+(x−xn)∆n (24)770

where the slope ∆n is determined according to the followingcases. If cn−1≤ cn≤ cn+1 or cn−1≥ cn≥ cn+1 then:

∆n = sign(cn+1−cn−1)×

min(cn+1−cn−1

2∆x,cn+1−cn

∆x,cn−cn−1

∆x

)(25)

where, in this case, ∆n is the smallest slope that can be775

estimated between cell n and its closest neighbours.Otherwise, cn is an extremum of concentration and the

concentration within cell n is assumed constant (∆n = 0)which ensures that the scheme is monotonic.

This scheme, which is recognized in meteorology for its780

good numerical accuracy and smaller diffusion than the first-order upwind scheme, is also slightly more time-consumingthan the first-order upwind scheme. In meteorology, it can beconsidered as a good compromise solution between numer-ical accuracy and computational efficiency for long-range785

transport (Hourdin and Armengaud (1999)).

5.1.3 The Parabolic Piecewise Method scheme

Another scheme that is proposed in CHIMERE for horizontaltransport is the 3rd order Parabolic Piecewise method (PPM)scheme, with slope-limiting and monotonicity-preserving790

conditions, such as presented in Colella and Woodward(1984). This scheme is applied for each dimension sepa-rately, which formally limits the model to 2d order accu-racy in solving the two-dimensional transport problem be-cause the cross derivative ∂c/∂x∂y is not taken into account.795

However, since the scheme is symmetric, it can be consid-ered that the errors due the neglect of cross-derivatives ap-proximately compensate (Ullrich et al. (2010)). Therefore,since the treatment of transport in each horizontal dimensionhas 3rd order accuracy, the PPM scheme as implemented in800

CHIMERE is much less diffusive than the simple 2d orderVan Leer scheme (see, e.g. Vuolo et al. (2009b)).

5.1.4 Comparison between the three available transportschemes

Vuolo et al. (2009b) have performed a comparison between805

the three horizontal transport schemes presented above in thecontext of an event of long-range transport of saharian dustover Europe. They found that the choice of one or another ofthese transport schemes has a strong impact on modelled dust


concentrations. As can be expected from the well-known nu-810

merical behaviour of these schemes, simulated peak valuesof the dust plume are reduced by 32% (upwind) or 17% (VanLeer) compared to the less diffusive PPM scheme, while theplume area, defined as the surface around the peak wheredust concentration exceeds 40% of the peak value, is in-815

creased by 48% (upwind) or 25% (Van Leer) compared to thePPM scheme. Horizontal transport schemes also have an in-direct impact on vertical transport and diffusion, the most dif-fusive horizontal schemes tending to increase dust transporttowards the lowest layers, increasing there domain-averaged820

surface concentrations and decreasing domain-averaged con-centrations in and above the boundary layer. The authorsconclude that the modelling uncertainty due to the choice ofone or another numerical transport scheme is among the lim-iting factors for the use of dust-transport models for opera-825

tional air-quality monitoring over Europe.

5.2 Vertical transport

5.2.1 Explicit vertical transport

In the vertical direction, unless otherwise specified by theuser the thickness of the layers increases quasi-exponentially830

with altitude in order to provide a better vertical resolutionin the lower model levels. The value of the thickness ra-tio between two neighbouring layers is typically close to1.5. Therefore the hypothesis of constant length for the gridcells cannot be made for vertical transport in CHIMERE, and835

other numerical schemes have to be used.First, vertical mass fluxes are calculated to secure zero flux

divergence at each grid cell. The vertical mass flux at thelower boundary of the lowest layer is zero, and the verticalmass flux at the top of each grid cell is computed succes-840

sively, from the lowest layer to the highest.Once these mass fluxes are known, the vertical transport

scheme can be applied. As the number of vertical layersin CHIMERE is much lower than the horizontal size of thedomain (typically 8-15 vertical layers, while horizontal do-845

mains have typically at least 40x40 grid cells, it is not clearwhether using high-order transport schemes relying on theconcentration values of several neighbouring cells is use-ful for vertical transport. Therefore, historically, only theclassical upwind scheme was used for vertical transport in850

CHIMERE, with the same formulation as presented abovefor horizontal transport.

However, more recent applications concern long-rangetransport of species having long lifetimes (e.g. mineral dust,volcanic ashes, particulate matter from forest fires) which855

also can occur above the boundary layer. Therefore, it seemsimportant to reduce numerical diffusion also in the verticaldirection. This is particularly important since numerical dif-fusion reduces the ability of the model to adequately repre-sent dense plumes that are located in thin vertical layers such860

as mineral dust or volcanic ashes. Therefore, a current devel-

opment in CHIMERE is to include the Van Leer I transportscheme, which is less diffusive than the upwind scheme, in aversion adapted to grid cells with nonuniform thickness. Thisscheme has been tested with encouraging results in terms865

of preserving sharp concentration gradients and higher peakvalues than the upwind scheme during long-range transportand shall therefore be proposed to the community in the nextdistributed version of CHIMERE.

5.2.2 Turbulent mixing870

At each interface between layers k and k+1, an equivalentturbulent vertical velocity wk is calculated:

wk =Kz

12 (hk+hk+1)

(26)

This allows the net incoming flux at the upper interface ofcell k to be diagnosed as:875

Fi =wk

(ck+1

ρk

ρk+1−ck

)hk

(27)

where ck is the concentration at the kth layer, ρk the air den-sity, hk the thickness of the kth layer.

The profiles of Kz and wk are computed in the meteorog-ical diagnostic code at each coarse time step, while the flux880

for each species depending on its concentration is computedat each fine time step (see section 4.1).

6 Emissions

Emissions of pollutants have different origins and include anumber of different gaseous and aerosol species, chemically885

inert or not. The sources can be located at the surface (traf-fic, biogenic) or along vertical profiles (industrial emissions,biomass burning). These emissions are split in several fami-lies, representing their origin:

• The anthropogenic emissions include all human activities890

(traffic, industries, agriculture, among others). They maybe very specific and emissions inventories are often dedi-cated to local scale simulation domain. When operated inthe Paris area, CHIMERE uses the AIRPARIF air qualitynetwork inventory (Valari and Menut (2008), Valari and895

Menut (2010)). For European studies, the EMEP inven-tory is usually used (Menut et al. (2012)), but the TNO in-ventory has also been used (Kuenen et al. (2011)). In stud-ies over North America, the model is also used with theU.S. EPA inventory (Solazzo et al. (2012b)). Finally, the900

EDGAR global emissions inventory was recently addedin CHIMERE and comparisons with the EMEP inventoryover Europe are on-going.• The biogenic emissions represent activities related to the

vegetation. These emissions are computed using the905

global MEGAN model (Guenther et al. (2006)).


• The mineral dust emissions represent the other part of”natural” emissions, but for non-reactive particle mainlygenerated by the surface layer dynamics (Menut et al.(2009a)). They are specific of several regions (Western910

Africa, Sahel and Saudi Arabia) but are described glob-ally. Some other sources exist in Europe but are not yet im-plemented and this specific physics is an on-going projectin CHIMERE. This is the case of the resuspension, actu-ally parameterised following the Loosmore and Cederwall915

(2004) scheme.• The fire emissions are more sporadic and require numer-

ous and very different data: satellite to estimate the burnedarea each day, a vegetation model to estimate the emit-ted amount for each chemical species (gas and particles).920

These emissions are also a specific development project inCHIMERE.

6.1 Anthropogenic emissions

The anthropogenic emissions are key in pollution manage-ment, since they are the only sources we can reduce. Con-925

trarily to the dust and biogenic emissions (only dependent onthe surface types and the meteorology), the anthropogenicemissions have to be prepared in a bottom-up way, using anumber of input data and information to build up an inven-tory of fluxes of chemical species. These various input infor-930

mations are generally given for a reduced number of classesof chemical species and often provided under an activity sec-tors classification, eg. following the ”Selected Nomenclaturefor Air Pollutants” (SNAP) nomenclature. It consist in yearlymasses per surfaces for various domains and resolutions.935

For a realistic simulation, these emissions must be pro-vided every hour, for the specific species of the chemicalmechanism used and projected over the gridded domain, re-gardless of the original data projection. From the raw data tothe data required for a specific simulation, a sequence of pre-940

processing actions is necessary, including a temporal disag-gregation, the application of VOC, NOx and PM shares andthe final species lumping into model species. A standard pro-cedure is proposed but preprocessing actions can be bypassedby directly providing hourly anthropogenic emissions. For945

the standard procedure, it is proposed to prepare these datafollowing two distinct stages, as displayed in Figure 9 :

• Step 1: Create an annual total gridded database per activ-ity sector adapted to the horizontal simulation grid. Thisenables then to create monthly masses of emitted model950

species, already projected on the horizontal simulationgrid, using seasonal factors (provided only for Europe).This step is performed only once each time a new domainis used. The complete suite of programs is provided to theuser only for the EMEP format, but can be adapted to other955

formats.• Step 2: Disaggregate the monthly emissions into hourly

emissions by applying daily and weekly factors, and thenproduce hourly emission time series for each species

Fig. 9. General principle of the CHIMERE anthropogenic emis-sions procedure.

adapted to the specific simulation period (real days) and960

the model vertical grid. This step is performed for eachsimulation. The complete suite performing this secondstep is provided to the user and does not need to be modi-fied.

Instead of going through Step 1, the user can also provide965

monthly anthropogenic emissions files built by other means.In these files, soil NO has not to be provided being consid-ered as biogenic emissions (even if agricultural activities areanthropogenic). This leads to monthly files for each speciesand for a typical year. For the MELCHIOR chemical mecha-970

nism implemented by default in CHIMERE, emitted speciesare listed in Table 3 .

Depending on the spatial domain, CHIMERE has beenused with several anthropogenic emissions datasets. Thelargest number of studies was over western Europe and two975

datasets were used: (i) the EMEP database (Vestreng (2003))and, more recently, the TNO database during the GEMS andMACC projects (Kuenen et al. (2011)). These data are spa-tially interpolated to the model grid. This is performed us-ing an intermediate fine grid with a 1km resolution (GLCF980

dataset, Hansen and Reed (2000)). Soil types described onthe fine grid allow for a better apportionment of the emis-sions according to urban, rural, forest, crops and maritimeareas. This pre-processing is provided with the model distri-bution to all users.985


Model species NameNO Nitrogen monoxideNO2 Nitrogen dioxideHONO Nitrous acidSO2 Sulphur dioxideNH3 AmmoniacCO Carbon monoxideCH4 MethaneC2H6 EthaneNC4H10 n-ButaneC2H4 EtheneC2H6 EthaneC3H6 PropeneC5H8 IsopreneOXYL o-XyleneHCHO FormaldehydeCH3CHO AcetaldehydeCH3COE Methyl ethyl KetoneAPINEN α-pinenePPM fin Primary particulate matterPPM coa Primary particulate matterPPM big Primary particulate matterH2SO4 fin Primary sulphuric acidOCAR fin Primary organic carbonBCAR fin Primary black carbon (or elemental carbon)

Table 3. List of MELCHIOR anthropogenic emitted species.

The way to estimate these emissions could have a strongimpact on modelled concentrations. In Menut et al. (2012),a sensitivity study was done to quantify the impact of thetemporal profil used to disagregate emissions mass fluxes.More recently, continental scale modelling was done over990

the United States and the model used the USEPA inventoryto model air quality during the AQMEII project, Rao et al.(2011), Schere et al. (2012), Solazzo et al. (2012b). Sev-eral other applications were also performed over Mexico Cityin the framework of the MILAGRO project, Hodzic et al.995

(2009).At a finer scale, CHIMERE is used with more specific

anthropogenic emission inventories. This is the case of allstudies done over the Paris area, using the AIRPARIF airquality network data (from Menut et al. (2000a) to Valari1000

et al. (2011)). A sensitivity study was presented in Valariand Menut (2008) showing the impact of the emissions hor-izontal resolution on the modelled concentrations: the effectof spatially ’averaged’ areas may lead to large changes in theemission fluxes and therefore concentrations: due to nonlin-1005

earity in chemical regimes, the modelled concentrations arenot varying linearly with the NOx and VOCs emission fluxeschanges. As one of the strongest ’air pollution’ hotspot inEurope, the model is also used to simulate the Po Valley pol-lution with a specific inventory as described in de Meij et al.1010

(2009), among others.Whatever the database and its resolution, the species NOx

and NMVOCs have to be distributed into the chemical mech-anism species. Annual emissions of NOx are first speciatedas 9.2% of NO2, 0.8% of HONO and 90% of NO following1015

GENEMIS recommendations (Friedrich (2000), Kurtenbachet al. (2001), Aumont et al. (2003)). Emissions of SOx arespeciated as 99% of SO2 and 1% ofH2SO4. The GENEMISNMVOC speciation is used for the same districts and for 6types of emission activity sectors: traffic, solvents, industry1020

(except solvents), energy extraction/production, residential(except solvents) and agriculture. For each activity, a speci-ation is obtained over 32 NMVOC NAPAP classes (Middle-ton et al. (1990)). Once the disagregation step is performed,an aggregation step for the lumping of NMVOCs into model1025

species is achieved following Middleton et al. (1990).For the anthropogenic emissions of primary particles,

H2SO4, PPM, OCAR, BCAR are split over three modes:

• XXX fin for diameters Dp < 2.5 µm• XXX coa for 2.5 < Dp < 10 µm1030

• XXX big for Dp > 10 µm.

PPM fin reffers to PM2.5, PPM coa to PM10-PM2.5.H2SO4, OCAR, BCAR are assumed to be in the fine mode.In rural areas, NO emissions from ammonium used in fertil-izer application, followed by microbiological processes may1035

be significant. Since these emissions strongly depend ontemperature, they are processed in the model as ”biogenic”emissions and we refer to them as ”biogenic NO emissions”.CHIMERE uses an European inventory of soil NO emis-sions from Stohl et al. (1996). This inventory estimates the1040

soil emission to be of the order of 20% of the emissionsfrom combustion on a European average, during the sum-mer months, but with large difference between the countries.In the model, these NO emissions are only considered duringthe months of May to August.1045

6.2 Biogenic emissions

Emissions of six CHIMERE species: isoprene, α-pinene, β-pinene, limonene, ocimene, and NO, are calculated using theMEGAN model data and parameterisations. The MEGANmodel (Guenther et al. (2006), v.2.04) exploits most recent1050

measurements in a gridded and canopy scale approach, moreappropriate for use in CTMs since it estimates the effectiveburden of gases that mix and react in the boundary layer. Es-timates of biogenic VOCs from vegetation and NO emissionsare calculated as :1055

ERi =EFi×γi(T,PPFD,LAI)×ρi (28)

where ERi (µg.m−2.h−1) is the emission rate of speciesi, EFi (µgm−2h−1) is an emission factor at canopy standardconditions, γi (unitless) is an emission activity factor thataccounts for deviations from canopy standard conditions, and1060

ρi is a factor that accounts for production/loss within canopy.


As a first step, canopy standard conditions are set to: airtemperature (T) of 303 K, photosynthetic photon flux den-sity (PPFD) of 1500 µmolm−2s−1 at the top of the canopy,leaf area index (LAI) of 5 m2m−2 and a canopy with 80%1065

mature, 10% growing and 10% old foliage.The MEGAN model parameterises the bulk effect

of changing environmental conditions using three time-dependent input variables specified at top of the canopy: tem-perature (T), radiation (PPFD), and foliage density (LAI).1070

The production/loss term within canopy is assumed to beunity (ρ= 1). The equation can then be expanded as :

ERi =EFi×γT,i×γPPFD×γLAI (29)

The MEGAN model provides inputEF and LAI data overa global grid, hereafter projected on the CHIMERE model1075

grid. The current available choice for EF s is restricted tofollowing species: isoprene, α-pinene, β-pinene, myrcene,sabinene, limonene, δ3-carene, ocimene, and nitrogen oxide.NO biogenic emissions include contribution from both for-est and agricultural (fertilizers) soils. EF ’s are static and1080

refer to years 2000-2001. They are obtained summing upover several plant functional types (e.g. broadleaf and nee-dle trees, shrubs, etc...). LAI database is given as a monthlymean product derived from MODIS observations, referred tobase year 2000. Hourly emissions are calculated using 2-m1085

temperature and short-wave radiation from a meteorologicalmodel output. Terpene and humulene emissions are not cal-culated in this model version and are set to zero.

For European studies with CHIMERE, a comparison ofthe simulated formaldehyde column was presented in Curci1090

et al. (2010). Formaldehyde concentrations variability is pri-marily driven by the oxidation of biogenic isoprene overEurope. By comparison to satellite based observations(Aura/OMI), it was shown that MEGAN isoprene emissionsmight be 40% and 20% too high over the Balkans and South-1095

ern Germany, respectively, and 20% too low over IberianPeninsula, Greece and Italy (Curci et al. (2010)).

6.3 Sea salt emissions

In the same way as biogenic emissions, the sea salt emissionsneed to know the meteorological parameters at an hourly1100

time step and over the whole simulated period and domain.They are calculated by Monahan (1986) :

dF

dr= 1.373U3.41

10 r−3(1+0.057r1.05)101.19e−B2

(30)

B=0.38− log(r)

0.65(31)

F is the flux of sea salt particle number expressed in par-1105

ticles m−2 s−1 µm−1, r the particle radius in µm and U10 isthe wind speed at 10 m in m s−1.

6.4 Mineral particles emissions

In this model version, CHIMERE is able to make calcula-tions on regional areas anywhere on Earth. The most impor-1110

tant source of particulate matter is mineral dust. In the north-ern hemisphere, mineral dust is mainly emitted in Africa (Sa-hara and Sahel), when some emissions are also observed overland such as Europe. These emissions are sporadic but in-tense. In order to correctly model the total budget of par-1115

ticulate matter, mineral dust emissions are diagnosed in themodel. The goal of the mineral dust modelling is twofold:improve our understanding on this physical problem (emis-sions, transported thin layers) and, after long-range transport,estimate the relative part of mineral dust in the total bud-1120

get of aerosols near the surface (and thus accounted for airquality in Europe, for example), as displayed for example inFigure 10 .

The calculation of mineral dust emissions is splitted intotwo parts: (i) over the western Africa, (ii) over Europe.1125

Over western Africa, the analysis and forecast of mineraldust was primarily done in a different branch of CHIMEREcalled CHIMERE-dust. The developments of the dust emis-sions and transport are still ongoing but the development ofCHIMERE-dust was frozen in 2010 and all dust calculations1130

are now integrated in the current CHIMERE model, ensuringmore homogeneous developments.

The dust emission fluxes are calculated using the parame-terisation of Marticorena and Bergametti (1995) for saltationand the dust production model (DPM) proposed by Alfaro1135

and Gomes (2001) for sandblasting. In order to have a betteraccuracy and a lower computational cost, the DPM is opti-mised as presented in Menut et al. (2005b). Before calcu-lating the fluxes, the threshold friction velocity is estimatedfollowing the Shao and Lu (2000) scheme. A complete de-1140

scription of the dust calculation is presented in Menut et al.(2007).

For long-range transport simulations, the modelled do-main is very large and must include at the same time Africa(for emissions) and Europe (for the long-range transport and1145

deposition). This leads to a coarse horizontal resolution of1o×1o in many studies. In order to take into account the sub-grid scale variability of observed winds, the dust emissionsare thus estimated using a Weibull distribution for the windspeed, following Cakmur et al. (2004) and Pryor et al. (2005).1150

An extension of the African dust emission scheme wasdone for Europe to model a huge dust event in Ukraine,Bessagnet et al. (2008). This shows that it is possible tomodel local European erosion and retrieve an extreme eventof particles, such as those observed in the northern-western1155

Europe (Netherlands, Belgium).In Menut (2008), the impact of the meteorological forcing

(NCEP or ECMWF) on the dust emission fluxes was quanti-fied. During the studied period, two major dust events wereobserved. It was shown that the meteorological models are1160

able to diagnose wind speed values high enough to provoke


Fig. 10. Example of dust plumes observed and simulated for theFebruary, 24 and March, 8 2006 with CHIMERE. Satellite imagesare from NASA/MODIS.

dust saltation for one observed event, and not the other. Eachmodel diagnosing one of the two observed events but not thesame. This highlights the huge sensitivity of dust emissionsto the surface meteorology used. In Vuolo et al. (2009a), the1165

model results were compared to CALIOP lidar data and thevertical diffusion was quantified. In Menut et al. (2009a),an intensive observation period of the AMMA program wasmodelled in forecast mode to study the variability of the pre-dictability of modelled surface dust concentrations. It was1170

shown that the sum of all model uncertainties (emissions,transport, deposition) and of the spread of the forecasted me-teorology induces a variability in surface concentrations stillhigher than the required precision for European air qualityforecast. A sensitivity study was presented in Menut et al.1175

(2009b) and it was shown that a very small area in the Sa-haran desert (around the position of the former French nu-clear tests site during the 60’s) may explain the sporadic(but low) radionuclides concentrations measured sometimes

in the South of France. Finally, the dust emissions are only1180

calculated using a bottom-up approach and future interestingdevelopments could be to merge these calculations to satel-lite data to improve the calculated emitted dust flux (as inHuneeus et al. (2012) at the global scale).

Over western Europe, the erosion process is less impor-1185

tant than over Africa but have to be taken into account forthe local budget of air quality. Saltation is not the only natu-ral aerosol upward entrainment process: close to the surface,turbulence induces resuspension of freshly deposited smallparticles. The extraction results from the imbalance between1190

adhesive and lifting forces. Such particles can originate fromthe atmosphere or the biosphere, and are particularly easyto extract shortly after deposition (Loosmore and Cederwall(2004)). In order to represent these processes, we use a bulkformulation based on the simple resuspension rate empiri-1195

cal formula of Loosmore and Cederwall (2004), which wasshown to provide a very good fit to the available resuspensionmeasurement data. The particles are first deposited then re-suspended. In reality, deposition and resuspension are simul-taneous, and the available dust concentration on the ground is1200

governed by resuspension, washout by runoff and absorptionby soil water, production by deposition and other biologicalor mechanical processes. The detail of all these processes isessentially unknown, and we assume here that the availableconcentration of dust does only depend on the wetness of the1205

surface, as fully described in Vautard et al. (2005). In thisempirical view, the resuspension flux is governed by:

F =Pf(w)u1.43∗ (32)

where f(w) is the soil moisture factor, given by:{w<wt : fw = 1w>wt : fw =

√1+1.21(100(w−wt))0.68

(33)1210

with w the gravimetric soil moisture (kg kg−1) and wt(kg kg−1) a threshold gravimetric soil moisture. An uniformvalue of 0.1 kg kg−1 is used, corresponding to a large clayfraction. P is a constant tuned in order to approximatelyclose the PM10 mass budget and the u∗ dependency follows1215

the Loosmore and Cederwall (2004) formulation. In absenceof any dedicated measurements, the reentrained PM10 parti-cle mass is distributed in a standard atmospheric size distri-bution: 2/3 of the mass as PM2.5 and 1/3 as coarse PM10-PM2.5. Within PM2.5, particles are distributed in the same1220

three modes as for the anthropogenic emissions.

6.5 Fire emissions

Fires are now recognized to be a major source of emissionsof aerosols and trace gases. Depending on the area studied,the species of interest and the time period analysed, it may1225

be necessary to account for this additional contribution in themodel simulations. Several studies have used CHIMERE to


evaluate the impact of large fire events on air quality at re-gional scale, for example Hodzic et al. (2007) for fires inPortugal in 2003 or Konovalov et al. (2011) for fires in Rus-1230

sia in 2010. A new emission preprocessor is currently be-ing developed in order to allow the evaluation of emissionseither from pre-existing emission inventories (e.g. van derWerf et al. (2010); Wiedinmyer et al. (2011)), or from thefire location points and estimated area burned directly. The1235

latter allows the construction of an emission inventory forthe CHIMERE grid and time-period chosen by the user. It isadaptable to near-real time observations for forecasting pur-poses.

This inventory is based on the general formulation of1240

Seiler and Crutzen (1980). For each model species i, theemission associated to a specific fire Ei (kg species) is esti-mated by multiplying the area burned in the correspondingvegetation type Aveg (m2) by the fuel load FLveg (kg drymatter (DM) m−2) and the specific emission factor EFi,veg1245

(g (kg DM)−1), as summarized in equation 34 :

Ei =nveg∑v=1

(AvegFLvegEFi,veg) (34)

The emissions are then binned into the specified modelgrid. The temporal and horizontal resolutions of the fireemissions depend on the resolution of the different parame-1250

ters. The area burned parameter is estimated from the globalobservations of fire activity and areas burned at a daily and1km resolution from the MODIS instrument (Giglio et al.(2010)), coupled to the SEVIRI/METEOSAT observationsRoberts and Wooster (2008) for the regions covered (Europe1255

and Africa) to allow the evaluation of a diurnal cycle. Anexample is presented in Figure 11 for May, 8 2012. De-pending on the fire location, a specific vegetation burned isattributed using the USGS landuse database (at 1km reso-lution) and the corresponding fuel load (or carbon content)1260

is evaluated from simulations by the ORCHIDEE vegetationand carbon cycle model (Krinner et al. (2005)).

Finally, the emissions are converted from carbon (or DM,considering that it is 45% carbon as in van der Werf et al.(2010)) to each species using emission factors from the Ak-1265

agi et al. (2011) review. Any species may be included in theinventory provided that emission factors are available. For afull description of each step of the calculation, the reader isreferred to Turquety (2012).

In addition to the amount of trace gases and aerosols, the1270

injection altitude is a critical parameter. Indeed, fires can re-lease enough energy to trigger or reinforce convection (Fre-itas et al. (2007); Rio et al. (2010)). These events will be ac-counted for by using a parameterisation of pyroconvection -still being implemented - to evaluate emission profiles based1275

on the fire intensity and the meteorological conditions (e.g.WRF simulations used for the CHIMERE simulations). Adatabase including all CHIMERE species will be made avail-able to users via the ECCAD portal (ECCAD) and the near-

Fig. 11. Example of area burned (km2) calculated using the fireemissions preprocessor for May, 8 2012 over western Europe.

real time evaluation of the emissions is on-going at LMD1280

(COSY (2013)). Updates on the availability of the fire emis-sion module will be available from the CHIMERE web page.

7 Chemistry

7.1 Chemical preprocessor

The complete chemical mechanism used by CHIMERE is1285

built using a suit of scripts and programs called chemprep.CHIMERE offers the option to include different gas phaseand aerosol chemical mechanisms. The originality inCHIMERE is that these chemical mechanisms are writtenin ASCII format, and, thus, do not need to be compiled.1290

The user can easily change some reactions or add new ones.When the model is launched, the availability of chemistry in-put files corresponding to the specific simulation is checked.If the data already exist, the run continues. If not, a pre-processor script will create the data directory. The strength of1295

this approach is that the user may easily create very particu-lar chemical schemes. All chemical parts being independent,the user may choose to have gas chemistry or not, aerosolschemistry or not (and the number of bins), sea salt, dust, sec-ondary organic aerosols, persistent organic pollutants, tracers1300

or not. The nomenclature was defined to be easily under-standable and the user can build chemistry sensitivity stud-ies by changing reactions rates, photo dissociation rates, etc.The list of active species may also be changed and chemicalfamilies may be defined (e.g. NOx=NO+NO2), so that they1305

can be dignosed as other tracers.

7.2 Gas phase chemistry

By default, there exist one aerosol scheme and two versionsof the gas phase scheme. The complete scheme, which is theoriginal scheme of Lattuati (1997), hereafter called MEL-1310

CHIOR1, describes more than 300 reactions of 80 gaseous


species. The hydrocarbon degradation is derived from thatof the EMEP gas phase mechanism (Simpson (1992)), withmodifications in particular for low NOx conditions and NOx-nitrate chemistry. All rate constants are taken from Atkinson1315

et al. (1997) and De Moore et al. (1994). Heterogeneous for-mation of HONO from deposition of NO2 on wet surfaces isconsidered, using the formulation of Aumont et al. (2003).For other heterogeneous reactions, see also section 7.3.3.

Fig. 12. Reaction pathways of the RO2 radical in MELCHIOR 1.

Inorganic chemistry (42 reactions) is treated in a classical1320

way, similarly to the original EMEP mechanism, and includ-ing the relevant chemistry of the tropospheric ozone-NOx-VOC system. Organic chemistry is based on the simplifieddegradation of 8 hydrocarbons and two alcohols. These com-pounds represent either individual species, generally for the1325

smallest molecules of a class (methane, ethane, ethene, iso-prene and methanol), or families of compounds (n-butanefor alkanes, propene for alkenes, o-xylene for aromatics, α-pinene for terpenes and ethanol for alcohols). These VOCsundergo oxidation reactions with OH, NO3, and ozone (the1330

latter only for the unsaturated compounds) leading to the for-mation of peroxy (RO2) radicals. All major reaction path-ways of the 25 RO2 radicals (including those formed by theoxidation of carbonyl compounds or nitrates, see below) arerepresented in the mechanism ( Figure 12 ):1335

• Reactions with NO leading to the formation of carbonylcompounds (including, when significant, the fragmenta-tion pathway)

• For some RO2 + NO reactions, a second pathway yieldingnitrates is taken into account1340

• RO2 reactions with NO3, important during night time, andresulting in the same VOC species than the RO2 + NOreaction

• RO2 reactions with NO2 resulting in the formation of per-oxynitrates (including PAN)1345

• Reaction with HO2 yielding hydroperoxides. Individualhydroperoxides (15) are taken into account. Their oxida-tion with OH or by photolysis is treated and yields the car-bonyls that also results from the RO2 + NO reaction.

• Recombination reactions of RO2 radicals. A full treat-1350

ment would require the treatment of 1/2(N2 +N) reac-tions (i.e. 325 for N=25) and would be too time consum-ing. This mechanism is simplified by taking into account

only the RO2 reaction with itself, with the most abundant(CH3O2) and the most reactive (CH3COO2)RO2 species.1355

Both the radical terminating and non-terminating recom-bination pathways are included (101 reactions).

Secondary VOC species formed from these reactions arecarbonyl compounds (9), hydroperoxides (15), nitrates (9)and peroxynitrates (4). As the primary VOCs, they can un-1360

dergo reactions with OH, NO3 and O3 in addition to photol-ysis (only for oxidized VOCs). A 0D study conducted byDufour et al. (2009) under low and high NOx conditionsallowed the comparison of formaldehyde yields from 10organic compounds, simulated by MELCHIOR1 and three1365

reference mechanisms (Master Chemical Mechanism fromSaunders et al. (2003), the SAPRC99 scheme developed byCarter (2000) and the fully explicit self-generated chemicalscheme SGMM of Aumont et al. (2005)). The results showthat MELCHIOR1 simulated yields agree within 20% with1370

the reference mechanisms. This agreement increases to 5%in high NOx conditions for C2H6, C3H6, CH3CHO, n-C4H10

and CH3OH oxidation.In order to reduce the computing time a reduced mecha-

nism with a bit more than 40 species and about 120 reactions1375

is derived from MELCHIOR1 (Derognat et al. (2003)). Thisscheme (MELCHIOR2) is optimised for polluted conditions.The concept of ”chemical operators” (Carter (1990)) hasbeen introduced, where RO2 radicals are treated as virtualspecies independently of their organic rest R. VOC degrada-1380

tion results in secondary compounds as if the correspondingRO2 radical reacted with NO. In our scheme, some individ-ual RO2 radicals are explicitly taken into account (CH3O2,CH3CO(O2), C5H8(OH)O2). As a further reduction step,minor reaction pathways under polluted conditions are ne-1385

glected. Under polluted and moderately polluted conditions(NOx > 100 ppt), differences between the reduced and thecomplete mechanism are below 5% for ozone, below 10%for NOx and HOx, and below 20% for OH. The full list ofspecies and reactions of MELCHIOR2 are given in Appendix1390

A.Photolysis rates are calculated under clear sky conditions

as a function of height using the Tropospheric Ultraviolet andVisible (TUV) radiation model (Madronich et al. (1998)).Clouds are taken into account in a highly parameterised fash-1395

ion, where clear sky photolysis rates are multiplied through-out model columns by an attenuation coefficient A depend-ing on the total Cloud Optical Depth (COD). Three optionsfor the calculation of COD are available in CHIMERE, thusallowing to fit several meteorological forcings. A future ver-1400

sion of CHIMERE will include an on-line version of TUV totake into account hourly variations of aerosols concentrationsand their impact of photolysis rates.

A comparison of MECHIOR2 and SAPRC07 (Carter(2010)) for the production of secondary organic gaseous1405

species within CHIMERE was recently conducted for Eu-rope over a whole summer season in year 2005, with a res-


MELCHIOR2 SAPRC07R 0.71 ± 0.08 0.71 ± 0.08RMSE (ppbv) 13.69 ±2.14 13.18 ±2.03Bias (ppbv) 9.29 ±2.65 8.19 ±2.65

Table 4. Averaged values, with their standard deviation, of the cor-relation coefficient, the root mean square error and the bias, calcu-lated for the comparison of simulated ozone with Air Base data overEurope for summer 2005, using MELCHIOR2 (left) and SAPRC07(right).

olution of 0.16o. For this purpose, SAPRC07 was imple-mented in CHIMERE, together with a new aggregation ta-ble for anthropogenic emitted species and a specific prepro-1410

cessing of boundary conditions in order to fit the SAPRClumped species. Also, an up-to-date photolysis rate table, us-ing recent data from Sander et al. (2006) and IUPAC (2006,http://www.iupac.org/) was provided for both schemes us-ing the most recent version of the TUV model. The results1415

for ozone show quite comparable correlation coefficients,RMSE and bias (see Table 4 ) with a slight tendency forSAPRC07 to reduce the averaged overestimation of ozone,compared with a set of 1300 Air Base measurement stations.A latitude-height cross section of a city plume also showed1420

that the two mechanisms produce similar quantities of HOxradicals (albeit somewhat lower with SAPRC due to reducedozone production), the main difference being the speciationof organic nitrogen (approximately 25% of oxidized NOxspecies in both schemes), which is more in favor of PAN1425

species when using MELCHIOR2 while SAPRC producesa larger amount of organic nitrates RNO3 (10% against 5%with MELCHIOR2). This may impact the geographical ex-tent of ozone production. The possibility to select eitherSAPRC07 or MELCHIOR will be given in the next version1430

of CHIMERE.

7.3 Aerosol module

7.3.1 Aerosols size distribution

CHIMERE contains a sectional aerosol module which in-cludes emitted Total Primary Particulate Matter (TPPM),1435

secondary species such as nitrate, sulphate, ammonium, sec-ondary organic aerosol. In addition, natural dust can be sim-ulated, as well as sea salt aerosols either as passive tracers, oras interactive species that are in equilibrium with other ions.The organic matter and elemental carbon can be speciated if1440

their emission inventory is available. In this case, the totalprimary particulate matter is composed of OCAR + BCAR +PPM, where PPM represents the remaining unspecified partof primary species.

Sulphate is formed by SO2 oxidation through both gaseous1445

and aqueous phase pathways. Nitric acid is produced in thegas phase by NOx oxidation. N2O5 is converted to nitricacid via heterogeneous pathways by oxidation on aqueous

aerosols. Ammonia is a primary emitted base converted toammonium in the aerosol phase by neutralisation of nitric1450

and sulphuric acids. Ammonia, ammonium, nitrate and sul-phate exist in aqueous, gaseous and particulate phases in themodel. As an example, in the particulate phase the modelspecies pNH3 represents an equivalent ammonium as thesum of NH+

4 ion, NH3 liquid, NH4NO3 solid, and other salts1455

consistently with the ISORROPIA thermodynamic equilib-rium model (Nenes et al. (1998), see below).

Atmospheric aerosols are represented by their size distri-butions and chemical compositions (Bessagnet et al. (2005)).The sectional representation described by Gelbard and Sein-1460

feld (1980) has been used for the density distribution func-tion. The sectional approach is quite useful to solve the gov-erning equation for multicomponent aerosols. It discretizesthe density distribution function in a finite number of sizesections (Warren (1986)) so that all particles in section l have1465

the same composition and are characterized by their mass-median diameter Dp.

The discretisation of the density distribution function q fora given aerosol component, follows equation 35

q(x) =dQ

dx(35)1470

where x is the logarithm of the massm of the particle (x=ln(m)) and Q is the mass concentration function.Qkl (µg m−3) is the mass concentration of the kth aerosol

component within the size section l. The total mass concen-tration in the size section l is given by equation 36 :1475

Ql =∫ xl

xl−1

q(x)dx=∑k

Qkl (36)

The range of the discretized size distribution and the num-ber of size sections (nb) are both user defined. The de-fault range of the distribution is set to 40nm-10µm. A goodcompromise between numerical accuracy and computational1480

time is nb=8, as used in the PREVAIR system, Rouıl et al.(2009) with the following mass-median diameter intervals:Dp=0.039, 0.078, 0.156, 0.312, 0.625, 1.25, 2.5, 5, 10 µm.

7.3.2 Aerosols dynamics

Coagulation1485

Coagulation is modelled following the classical theory de-scribed in Gelbard and Seinfeld (1980). Considering thatQkl is the mass concentration of component k in size sec-tion l, the mass balance equation for coagulation followsequation 37 :1490

[dQkldt

]coag

=12

l−1∑i=1

l−1∑j=1

[1aβi,j,lQ

kjQi+

1bβi,j,lQkiQj

](37)


−l−1∑i=1

[2aβi,lQiQ

kl −2bβi,lQlQ

ki

]− 1

23βl,lQlQ

kl −Qkl

m∑i=l+1

4βi,lQi

The sectional coagulation coefficients1aβ,1bβ,2aβ,2bβ,3β and 4β depend on particle charac-1495

teristics and meteorological data such as temperature,pressure and turbulence parameters (Fuchs (1964)). Forsubmicronic particles, coagulation is essentially driven byBrownian motions. The four terms on the right hand side ofequation 37 represent, respectively: the gain in section l1500

due to coagulation of particles in sections lower than l, theloss in section l due to coagulation of particles in section lwith those in sections lower than l, the loss by intrasectionalcoagulation in section l, and the loss due to coagulation ofparticles in section l with those in sections higher than l.1505

Gas-particle conversion

The implementation of the absorption process in CHIMEREis based on Bowman et al. (1997). The absorption flux J(µgm−3s−1) of species onto a monodisperse aerosol is:

J =1τ

(G−Geq) (38)1510

with G and Geq (µgm−3) the gas phase and equilibriumconcentrations respectively. The characteristic time τ is:

τ =1+

8λαDp

2πλcDpN(39)

with λ (m) the mean free path of air molecules, Dp (m)the diameter of the particles, N (particles m−3) the number1515

concentration, α the accommodation coefficient of the trans-ferred species and c (m s−1) the mean molecular velocity.For a semi-volatile species k, a mean absorption coefficientHkl (s−1) is defined at section l as:[dQkldl

]abso

=Hkl Ql (40)1520

Hkl =

12λckρpD2

p(1+(8λ/αkDp))(Gk−Gkl,eq) (41)

where ρp is the particle density (fixed at 1.5 gcm−3 here).Different absorption modules are implemented in

CHIMERE for the inorganic and organic aerosols. Forinorganic species (sulphate, nitrate, ammonium), the1525

equilibrium concentration Geq is calculated using thethermodynamic module ISORROPIA (Nenes et al. (1998)).This model also determines the water content of particles.Chloride and sodium can be optionally included with asignificant increase in computational time. The model1530

calculates the thermodynamical equilibrium of the sul-phate/nitrate/ammonium/sodium/chloride/water system ata given temperature and relative humidity. The possiblespecies for each phase are the following:

• Gas phase: NH3, HNO3, HCl, H2O.1535

• Liquid phase: NH+4 , Na+, H+, Cl−, NO−3 , SO2−

4 , HSO−4 ,OH−, H2O, HNO3(aq), HCl(aq), NH3(aq), H2SO4(aq).• Solid phase: (NH4)2SO4, NH4HSO4, (NH4)3H(SO4)2,

NH4NO3, NH4Cl, NaCl, NaNO3, NaHSO4, Na2SO4

Due to its low vapour pressure, sulphuric acid is assumed1540

to reside completely in the condensed phase. Sodium is al-ways in the liquid or solid phases. The solid/liquid phasetransition is solved with ISORROPIA by computing the del-iquescence relative humidities (relative humidity at the tran-sition point between the two phases).1545

Two ways are possible:

• Case 1: Reading a look-up table already prepared us-ing ISORROPIA (and provided with the model input datafiles)• Case 2: Using the implemented on-line coupling of ISOR-1550

ROPIA in CHIMERE.

In case 1, the calculation can be done by interpolating apre-calculated look-up table ( Table 5 ). The partitioning co-efficient for nitrates, ammonium and the aerosol water con-tent has been calculated for a range of temperatures from 2601555

to 312K, relative humidities from 3% to 99% and concentra-tion ranges from 10−2 to 65 µgm−3. Because of numericallimitations, sodium and chloride are not accounted for in thistable. In case of the active ”sea salt” option, the model auto-matically switch to case 2.1560

Variable Value IncrementMin. Max.

Temperature (K) 260 312 + 2.5Relative Humidity 0.3 0.99 + 0.05H2SO4, HNO3, NH3 (µgm−3) 10−2 65 × 1,5

Table 5. Look-up table used for the calculation of the thermody-namic equilibrium with ISORROPIA. The minimum and maximumvalues represent the range of the values calculated. Their values aredefined to cover a large range of possible meteorological situations.In case of temperature, relative humidity or concentrations less thanthe minimum or up to the maximum, the thermodynamic equilibriumis chosen as the value corresponding to the last defined value (min-imum or maximum). The increment is defined to ensure a realisticlinearity between two consecutive values and for the interpolation.It may be an additive (+) or a multiplicative (×) increment.

The use of the look-up table allows to run the model faster,at the expense of accuracy around each deliquescent point.Comparisons with on-line coupling was done and presentedin Hodzic (2005), finding that the use of an on-line coupling


leads to a weak decrease of the mean concentrations (13%1565

for the nitrates, and no more than 2% for the sulphates). Am-monium concentrations are slightly increased (1 to 5%). Inabsolute values, the differences never exceed 0.5 µgm−3 inaverage for the nitrates and 0.1 µ g m−3 for the sulphates andammonium. The aerosol water content is slightly decreased1570

(3 to 13%, 0.1 to 2.8 µ g m−3).For semi-volatile organic species, the equilibrium con-

centration of the aerosol component k at the size section l(Gkl,eq) is related to the particle concentration Qkl through atemperature dependent partition coefficient Kp (in m3µ g−1)1575

(Pankow (1994)):

Gkl,eq =Qkl

OMlKpk

(42)

with OM (µgm−3) the concentration of the absorptive or-ganic material. Considering the thermodynamic equilibriumbetween the gas and particulate phases, this coefficient is1580

given by

Kpk =

10−6RT

MWomζkp0k

(43)

with R the ideal gas constant (8.206 10−5m atmmol−1K−1), T the temperature (K), MWom the mean molec-ular weight (g mol−1), p0

i the vapour pressure of product i1585

as a pure liquid (atm) and ζ the activity coefficient of speciesin the bulk aerosol phase. The coefficient ζ is difficult tocalculate and is assumed constant and equal to one.

For the nucleation process of sulphuric acid, the param-eterisation of Kulmala et al. (1998) is used. This process,1590

favoured by cold humid atmospheric conditions, affects thenumber of ultrafine particles. The nucleated flux is addedto the smallest bin in the sectional distribution. Nucleationof condensable organic species has been clearly identified inmany experimental studies (Kavouras et al. (1998), however1595

there is no available parameterisation. Since the sulphuricacid nucleation process competes with absorption processes,the nucleation is expected to occur in low particle pollutedconditions.

7.3.3 Aerosols chemistry1600

Multiphase chemistry

Sulphate production in the aqueous-phase occurs from thefollowing reactions (Berge (1993); Hoffmann and Calvert(1985); Lee and Schwartz (1983)):

• SOaq2 +Oaq3 →SO2−41605

• HSO−3 +Oaq3 →SO2−4

• SO2−3 +Oaq3 →SO2−

4

• SOaq2 +H2Oaq2 →SO2−

4

• SOaq2 +NOaq2 →SO2−4

• SO2−3 (Fe3+)→SO2−

41610

• HSO−3 (Mn2+)→SO2−4

SO2, H2O2 and O3 in the aqueous phase are in equilibriumwith the concentrations in the gas phase. Moreover, aqueousSO2 is dissociated into HSO−3 and SO2−

3 . Catalyzed oxi-dation reactions of sulphur dioxide in aqueous droplets with1615

iron and manganese are considered, following Hoffmann andCalvert (1985) among others. Henry’s law coefficient andother aqueous equilibrium constants are used (Seinfeld andPandis (1997)). Sulphur chemistry is very pH sensitive. ThepH is calculated by solving the charge balance equation in1620

the aqueous phase as described in Bessagnet et al. (2004).To avoid large uncontrolled variations, the pH may vary onlybetween 4.5 and 6.0.

A few heterogeneous reactions are also considered. Nitricacid is produced on existing particles and fog droplets. Al-1625

though aerosol particles and cloud droplets represent a smallfraction of the atmosphere, it is well established that reac-tions involving gas species onto their surfaces may signifi-cantly contribute to atmospheric chemistry cycles. For ozonemodelling, Jacob (2000) recommends to include a minimal1630

set of reactions:

• HO2→ 0.5H2O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .γ = 0.2• NO3→HNO3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . γ = 0.001• NO2→ 0.5HNO3 + 0.5HONO . . . . . . . . . . . . γ = 0.0001• N2O5→ 2HNO3 . . . . . . . . . . . . . . . . . . . . . . . . γ = 0.01 - 11635

with γ the associated uptake coefficients are provided inHarrison and Kito (1990), and other references in Jacob(2000).

The first-order rate constant k for heterogeneous loss ofgases onto particles is given by:1640

k=∑l

(Dp(l)2Dg

+4νγ

)−1

Al (44)

with Dp, the particle diameter (m), Dg , the reacting gasmolecular diffusivity (m2 s−1), ν, the mean molecular veloc-ity (m s−1), Al, the total surface area in the particle bin l andγ the uptake coefficient of reactive species. The uptake co-1645

efficient for N2O5 is assumed to be temperature-dependentin the range 0.01 - 1 (De Moore et al. (1994)) with in-creasing values for decreasing temperatures. Aumont et al.(2003) suggest that NO2 reactions on wet surfaces could bean important source for HONO production during wintertime1650

smog episodes, which is included in CHIMERE (also presentin the gas-phase mechanism, see above).

Secondary organic aerosol chemistry

The complete chemical scheme for SOA formation im-plemented in CHIMERE includes biogenic and anthro-1655

pogenic precursors ( Table 6 ) as described in Bessagnetet al. (2009). Biogenic precursors include API (α-pinene andsabinene), BPI (β-pinene and δ3-carene), LIM (limonene),


OCI (myrcene and ocimene) and ISO (isoprene). Anthro-pogenic precursors include TOL (benzene, toluene and other1660

mono-substituted aromatics), TMB (Trimethylbenzene andother poly-substituted aromatics), and NC4H10 (higher alka-nes). SOA formation is represented according to a single-step oxidation of the relevant precursors and gas-particle par-titioning of the condensable oxidation products. The gas-1665

particle partitioning formulation has been described in detailby Pun et al. (2006). The overall approach consists in dif-ferentiating between hydrophilic SOA that are most likelyto dissolve into aqueous inorganic particles and hydropho-bic SOA that are most likely to absorb into organic parti-1670

cles. The dissolution of hydrophilic SOA is governed byHenry’s law whereas the absorption of hydrophobic particlesis governed by Raoult’s law. The large number of condens-able organic compounds is represented by a set of surrogatecompounds that cover the range of physico-chemical prop-1675

erties relevant for aerosol formation, i.e., water solubilityand acid dissociation for hydrophilic compounds and satu-ration vapour pressure for hydrophobic compounds. Thesesurrogate compounds were selected by grouping identifiedparticulate-phase molecular products with similar proper-1680

ties. The molecular weight of each surrogate compoundis determined based on its structure and functional groups.The Henry’s law constant or the saturation vapour pres-sure of the surrogate species is derived from the averageproperties of the group. Other properties are estimated us-1685

ing the structure of each surrogate compound. Enthalpyof vapourisation are given in brackets (kJmol−1) for eachSOA coumpounds : AnA0D (88), AnA1D(88), AnA2D(88),BiA0D(88), BiA1D(88), BiA2D(109), AnBmP(88), An-BlP(88), BiBmP(175). The full name of compounds are ex-1690

plicited in Table 6 caption.The base SOA module was tested against the smog cham-

ber data of Odum et al. (1997) for anthropogenic compoundsand those of Griffin et al. (1999) for biogenic compoundsand was shown to satisfactorily reproduce SOA formation for1695

those compounds, Pun et al. (2006). Higher gaseous alkanesand isoprene were added to the original chemical mechanismof Pun et al. (2006). The formation of SOA from higher alka-nes follows the formulation of Zhang et al. (2007) for the sto-ichiometric SOA yield and it is assumed that the SOA species1700

can be represented by a hydrophobic surrogate compoundwith a moderate saturation vapour pressure. The formationof SOA from the oxidation of isoprene by hydroxyl radicalsis represented with two surrogate products and follows theformulation of Kroll et al. (2006); Zhang et al. (2007).1705

The base SOA module described here was compared withthe Aerosol Mass Spectrometer measurements in MexicoCity during the MILAGRO-2006 field project Hodzic et al.(2009), showing a tendency to underpredict the results by 2-8 times the observed levels of SOA in the city as well as at1710

the regional scale downwind of the city. The model has beenupdated to include the SOA formation from primary organicvapours that has been proposed as an additional and impor-

tant source of SOA (Robinson et al. (2007)), and that hasallowed to significantly improve the comparison with mea-1715

surement in Mexico City (Hodzic et al. (2010b)). Additionalconstraints from measurements have been included throughthe calculation of the oxygen to carbon ratios, or non-fossilto fossil carbon amounts (Hodzic et al. (2010a)). The code isavailable upon request.1720

8 Dry deposition

The dry deposition process is commonly described througha resistance analogy (Wesely (1989)). These resistances arenot expressed in the same way, considering gas or particlesas displayed in Figure 13 .1725

Fig. 13. Main principle of dry deposition for gas and particles.For each model species, three resistances have to be estimated: thedeposition is finally done if the sum of all these resistances is low.For particles, the settling velocity is added.

The deposition process is a sink and only acts on a con-centration c along the vertical, as:

∆c=∂

∂z(Vd.c) (45)

with Vd the dry deposition velocity. For each gaseous speciesor particles, the deposition is due to three different processes.1730

First, the turbulent diffusivity is needed to estimate the aero-dynamical resistance, ra. Second, the diffusivity near theground, in the ”laminar” layer, is needed to estimate thesurface resistance rb. Third, for gaseous species only, thespecies water solubility is needed to estimate the canopy re-1735

sistance rc. For particles, there is no solubility, the particleis subject to gravity, falling with a settling velocity vs. Thedry deposition velocity (in cm s−1) for gaseous species isexpressed as:

vd =1

ra+rb+rc(46)1740


Reactions kinetic rates (molec.cm−3.s−1)TOL+OH→ 0.004×AnA0D + 0.001×AnA1D + 0.084×AnBmP + 0.013×AnBlP 1.81×10−12exp(355/T)TMB+OH → 0.002×AnA0D + 0.002× AnA1D + 0.001×AnA2D + 0.088×AnBmP +0.006×AnBlP

9.80×10−9/T

NC4H10+OH→ 0.07×AnBmP 1.36×10−12exp(190/T)−2

API+OH→ 0.30×BiA0D + 0.17×BiA1D + 0.10×BiA2D 1.21×10−11exp(444/T)API+O3→ 0.18×BiA0D + 0.16×BiA1D + 0.05×BiA2D 1.01×10−15exp (-732/T)API+NO3→ 0.80×BiBmP 1.19×10−12exp(490/T)BPI+OH→ 0.07×BiA0D + 0.08×BiA1D + 0.06×BiA2D 2.38×10−11exp(357/T)BPI+O3→ 0.09×BiA0D + 0.13×BiA1D + 0.04×BiA2D 1.50×10−17

BPI+NO3→ 0.80×BiBmP 2.51×10−12

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

LIM+O3→ 0.09×BiA0D + 0.10×BiA1D 2×10−16

OCI+OH→ 0.70×BiA0D + 0.075×BiA1D 5.10×−8/TOCI+O3→ 0.50×BiA0D + 0.055×BiA1D 7.50×10−14/TOCI+NO3→ 0.70×BiA0D + 0.075×BiA1D 4.30×10−9/TISO+OH→ 0.232×ISOPA1 + 0.0288×ISOPA2 2.55×10−11exp(410/T)

Table 6. Gas phase chemical scheme for SOA formation in CHIMERE. The surrogate SOA compounds consist of six hydrophilic speciesthat include an anthropogenic nondissociative species (AnA0D), an anthropogenic once-dissociative species (AnA1D), an anthropogenictwice-dissociative species (AnA2D), a biogenic non dissociative species (BiA0D), a biogenic once-dissociative species (BiA1D) and a bio-genic twice-dissociative species (BiA2D), three hydrophobic species that include an anthropogenic species with moderate saturation vapourpressure (AnBmP), an anthropogenic species with low saturation vapour pressure (AnBlP) and a biogenic species with moderate saturationvapour pressure (BiBmP), and two surrogate compounds for the isoprene oxidation products.

and for particles as:

vd = vs+1

ra+rb+rarbvs(47)

8.1 The resistances

The aerodynamical resistance ra depends on several turbu-lent parameters such as L the Monin-Obukhov length, the1745

friction velocity u∗, the dynamical roughness length z0m.

ra =1ku∗

[ln

(z

z0

)−ΨM

( zL

)](48)

where ΨM is the similarity function accounting for surfacelayer stability, as defined in Zhang et al. (2001).

The quasi-laminary boundary layer resistance rb factor is1750

estimated as:

rb =2ν

k×DH2OwPrDH2O

2/3g (49)

with k the Karman number (here k=0.41), DH2Ow andDH2Og the molecular diffusivity of water and gaseousspecies, respectively, and Pr the Prandl number. For gaseous1755

species, the molecular diffusivity is expressed as:

DH2Og =

√dMx

18(50)

The main land/seasonal parameters follow seasonal vari-ations of resistances. Most land parameters are taken fromWesely (1989), but LAI are drawn from the NASA/EOSDIS1760

Species dMx dHx df0O3 48 0.01 1SO2 64 1e5 0NO2 46 0.01 0.1NO 30 2e-3 0NH3 17 1e+5 0

Table 7. The main characteristics of the dry deposited species withtheir names, dMx the molar mass of the model species, dHx theeffective Henry’s law constant (M atm−1) for the gas, df0 the nor-malised (0 to 1) reactivity factor for the dissolved gas.

Oak Ridge National Laboratory using average LAI field mea-surements for summer. The molar masses used in CHIMEREare displayed in Table 7 .

The formulation of the surface resistance rc follows Eris-man et al. (1994). It uses a number of different other resis-1765

tances accounting mainly for stomatal and surface processeswhich are again dependent on the land use type and season.Necessary chemical parameters for the calculation of rc arealso taken from Erisman et al. (1994) except for carbonyls(Sander et al. (1999); Baer and Nester (1992)) and perox-1770

ide species (Hall et al. (1999)). As dHx and df0, presentedin Table 7 , are used for the mesophyllic resistance value asdescribed in Seinfeld and Pandis (1997)). The mesophyllicresistance rm is calculated for each deposited species as:

Rm =1

dHx×3.310−4 +df0×102(51)1775


Over vegetal canopies, corrections have been implementedaccording to Zhang et al. (2001), Giorgi (1986) and Petersand Eiden (1992).

8.2 Settling velocity vs

The settling velocity represents the effect of gravity on parti-1780

cles. This velocity is expressed as:

vs =118D2pρpgCc

µ(52)

with ρp, the particle density (chosen as ρp=2.65 g.cm−3

for mineral dust),Dp the mean mass median diameter of par-ticles,Cc a slip correction factor accouting for the noncontin-1785

uum effects whenDp becomes smaller and of the same orderof magnitude than the mean free path of air, λ, Seinfeld andPandis (1997). g is the gravitational acceleration with g=9,81m/s2, µ the dynamic viscosity (here the air dynamic viscosityis set to µair = 1.8×10−5kgm−1s−1). The slip correction1790

factor Cc is estimated as:

Cc = 1+2λDp

[1.257+0.4exp

(−1.1Dp

2λ

)](53)

with λ the mean free path of air, estimated as:

λ=2µair

p

√8Mair

πRT

(54)

where Mair the molecular mass of dry air (here 28.81795

gmol−1), T the temperature (K), p the pressure (Pa), µ theair dynamic viscosity.

9 Impact of clouds

The clouds may impact the photolysis, the chemistry via dis-solution of gases in precipitating drops and wet scavenging.1800

9.1 Impact of clouds on photolysis

In this model version and only for the photolysis attenuation,clouds are assumed to lie above the model top, so that thereis no cloud albedo effect within the model domain depth.For all photolysed species, clear sky photolysis rates Jc(z)1805

are multiplied throughout model columns by an attenuationcoefficient A(d) depending on the total cloud optical depth(COD) d. Using the TUV model, and a large set of CODsfor clouds at various altitudes, the attenuation relative to theclear-sky case has been fitted as a function of COD with the1810

formula:

A(d) = e−0.11d2/3(55)

Several options are offered in order to calculate the COD.Total COD, d, is the sum of partial CODs from three cloud

layers, low clouds dl, medium clouds dm and high clouds dh.1815

The limits between these clouds is user-chosen, but dependon the meteorological model. For WRF or MM5, limits of2000m and 6000m are proposed. At the end, the model usesthe COD estimated for the related model vertical level. Foreach cloud layer, three options are possible for the calcula-1820

tion of the partial cloud optical depth. If the liquid and icewater content are available in the input meteorological fields,we recommend to use the first option.

• Calculation as a function of Liquid/Ice Water Content inthe column; assuming sphericity, equivalent droplet size1825

of 6 microns and hexagonal shape for ice particles, theformula for cloud optical depth is 180.Cw + 67.Ci, whereCw and Ci are respectively the liquid water column (in Kgm−2) and ice column for this cloud layer.• Another choice is a parameterisation using relative humid-1830

ity only. It consists in parameterising the COD as a func-tion of the integral R, over the cloud depth, of the relativehumidity above 75%. It is assumed that small cloud for-mation (in particular cumulus clouds) starts at 75% rela-tive humidity. Normalisation of R leads to the formulation1835

(for instance for low clouds): dl = aR/dz, where a=0.02 ischosen such that a 1000m-thick layer has an optical clouddepth of 20.• An even simpler parameterisation can be achieved by mak-

ing the COD simply proportional to the cloud fraction (if1840

available) for each cloud layer. Coefficient tuning led toproportionality coefficients of: 50 for low-clouds, 10 formedium clouds and 2 for high clouds. This means, for in-stance, that a sky covered with 100% of high clouds has anoptical depth of 2. Tuning was performed with ECMWF1845

cloud fraction data and should change with the meteoro-logical model.

9.2 Wet scavenging

Scavenging for gas/aerosols in clouds or rain droplets istaken into account as follows:1850

• For gases in clouds: Nitric acid and ammonia in the gasphase are scavenged by cloud droplets and this process isassumed to be reversible. During cloud dissipation, andfor a non precipitating cloud, dissolved gases may reap-pear in the gas phase (Bessagnet et al. (2004)). For a1855

gas denoted A and the processes between gas and aqueousphases, there are two simultaneous reactions occurring:

Ag→k+Aaq;Ag←k−Aaq (56)

The constants k− and k+ (s−1) are estimated followingthe relations:1860

k−=6wlρaρeD

(D

2DgA

+4

cAαA

)−1

(57)

k+ =600

RHAT

(D

2DgA

+4

cAαA

)−1

(58)


with ρe and ρa the water and air densities, respectively,in kgm−3, wl the liquid water content (kg/kg), D thedroplet mean diameter (m), cA the mean molecular veloc-1865

ity of the gasA (ms−1), Dg the molecular diffusion of thegas A in air (in m2s−1) and αA the gas A accomodationcoefficient. H is the Henry’s constant (Matm−1), T theair temperature (K) and R the molar gas constant.

• For gases in rain droplets below the clouds: Dissolu-1870

tion of gases in precipitating drops is assumed to be irre-versible, both for HNO3 and NH3. The scavenging coeffi-cient is expressed as:

Γ =pDg

6.105ugD2(2+0.6R1/2

e S1/3c ) (59)

p being the precipitation rate (mm h−1), Dg the molecular1875

diffusion coefficient (m2s−1), ug the raindrop velocity (ms−1), Re and Sc respectively the Reynolds and Schmidtnumbers of drops. Mircea and Stefan (1998) and refer-ences therein give relationships between ug and hydrome-teor diameter for various types of precipitation.1880

• For particles in clouds : Particles can be scavenged ei-ther by coagulation with cloud droplets or by precipitatingdrops. Particles also act as cloud condensation nuclei toform new droplets. This latter process of nucleation is themost efficient one in clouds. According to Tsyro (2002)1885

and Guelle et al. (1998), the deposition flux, Q, is writtenas:[dQkldt

]=−εlPr

wlhQkl (60)

where Pr is the precipitation rate released in the grid cell(g cm−2 s−1), wl the liquid water content (g cm−3), h the1890

cell thickness (cm) and ε an empirical uptake coefficient(in the range 0 - 1) depending on particle composition. land k are respectively the bin and composition subscripts.

• For particles in rain droplets below the clouds : Parti-cles are scavenged by raining drops, the deposition flux of1895

particles being:[dQkldt

]=−αpEl

ugQkl (61)

with α is an empirical coefficient, p the precipitation ratein the grid cell (g cm−2 s−1), E a collision efficiency co-efficient between particles and raining drops (Slinn (1983)1900

and ug the falling drop velocity (cm s−1). Assuming aconstant drop diameter (2 mm), this parameterisation is anapproximation of equations described in Seinfeld and Pan-dis (1997) and Jung et al. (2002).

10 Model results evaluation1905

CHIMERE integrates a large set of complex processes anddelivers chemical concentrations. These chemical concentra-tions have to be compared to available measurements in order

to: (i) understand complex and non-linear processes and in-vestigate geophysical hypotheses; (ii) validate the model for1910

specific chemical species and location against measurementsand therefore estimate the realism of the parameterisations inthe model; (iii) make sensitivity and scenario studies in or-der to quantify changes in emissions or meteorology underclimate change etc.1915

For all these reasons, the comparisons between modelledconcentrations and measurements were always carefully per-formed during the various stages of the model developments.These comparisons were carried out with three different op-tions:1920

• Improve the processes:The comparisons to field campaings measurements allowan assessment of the model ability to reproduce specificair quality episodes using a wide range of available mea-surements;1925

• Dynamic evaluation:Long-term simulations can be used to evaluate the sen-sitivity of the model to changing situations (i.e. monthlyvariability, seasonal cycles, long term trends in emissions);• Compare to other numerical tools:1930

Model inter-comparison exercices allow comparing theperformances of the model with respect to other state ofthe art models and developing ensemble approaches.

10.1 International projects

The Table 8 summarizes some of the international projects1935

where CHIMERE was involved. Some of them include large-scale field campaigns, with data measurements during Inten-sive Observations Periods that were used to understand pol-lution events, for the model development and its validation.Many other national projects (French, Spanish, Italian etc.)1940

were conducted during the last years and are not listed here.Some of these projects include field campaigns which

comprise a wide range of measurements during selected pol-lution events. The measurements are limited in space (a re-gion) and time (several days) but they include a larger set1945

of physical and chemical parameters than basic monitoringnetworks.

The first field campaign using CHIMERE was ESQUIFduring the summers 1998 and 1999. In this first version, themodel was a box model: five boxes with a central one repre-1950

senting the Paris city and only two vertical levels (the surfaceand the boundary layer), Menut et al. (2000a). Even if thatmodel version was very simple, the model was used in fore-cast mode and was of great help to choose the best periodsto launch the intensive observation periods (IOPs), more par-1955

ticularly for airborne measurements, Menut et al. (2000b),Vautard et al. (2003). During ESQUIF, the relative part oflocal ozone production and long-range transport was quanti-fied and it was shown that Paris pollution episodes can notbe above the legal limits with local production only. For the1960


first time, the predicted aerosol chemical and optical proper-ties were evaluated against chemically speciated aerosol dataand lidar measurements (Hodzic et al. (2006b)). In 2001,CHIMERE was also used for forecast and analysis duringthe ESCOMPTE campaign that was also devoted to photo-1965

oxidant pollution (ozone) but in the Marseille area, in south-ern France (Menut et al. (2005a). Between ESQUIF andESCOMPTE, the horizontal grid became cartesian, Schmidtet al. (2001) and covered the lower troposphere with 8 levels.

For the GEMS project, the model was spatially extended1970

to the whole Western Europe, switching the anthropogenicsource to the EMEP emissions. These emissions have beenchanged for the MACC project and the TNO inventory wasimplemented, Zyryanov et al. (2012).

In 2006, mineral dust emissions and transport were added1975

to the model and used during the AMMA campaign. Themodel was used in forecast mode and the predictabilityof mineral dust emissions was quantified, Menut et al.(2009a). The model version with gaseous and aerosolsspecies was used during several projects such as MILAGRO,1980

over the Mexico area (Hodzic et al. (2009), Hodzic et al.(2010b), Hodzic et al. (2010a), Hodzic and Jimenez (2011));MEGAPOLI over the Paris area (Royer et al. (2011)),EC4MACS for the climate and air pollution mitigation strate-gies in Europe, AQMEII for model inter-comparisons be-1985

tween the United States and Europe (Pirovano et al. (2012)).More recently, two European projects have been the opportu-nity of new developments: the CIRCE project where the firston-line coupling between chemistry and vegetation was donebetween CHIMERE and ORCHIDEE, and ATOPICA for the1990

development of a new module for the pollen modelling.

10.2 Evaluation in extreme events

Air quality simulations are particularly relevant during ex-treme events in order to anticipate potentially detrimental sit-uations. CHIMERE was used to simulate air quality in sev-1995

eral such events. It was shown to accurately reproduce ob-served ozone concentrations during the 2003 European heatwave (Vautard et al. (2005)) and transport of aerosol smokeplumes (Hodzic et al. (2007)) across Europe during the 2003heat wave. The extreme particulate matter episode that took2000

place in Germany earlier in 2003 was also simulated but withless success by CHIMERE as well as other models, Sternet al. (2008).

In 2008, an unexpected event of high particulate matterconcentrations at the surface was observed in Belgium and2005

Netherlands. First thought to be long-range transport of min-eral dust from Africa, these huge concentrations were finallyidentified as coming from Ukraine. This assessment wasachieved by including a new dust mineral source represent-ing the erodible Chernozemic soil. The validation was done2010

with surface (AirBase) and space-borne lidar (Caliop) mea-surements, Bessagnet et al. (2008).

CHIMERE was used to simulate the transport of the plumeof the extreme fire incident that occurred in the Buncefieldoil depot in late 2005 (Vautard et al. (2007)). It was shown in2015

particular that the lack of major air quality degradation wasdue to the dispersion and transport of the enormous plume ofparticulate matter above the boundary layer. More recently,the dispersion of the Eyjafjallajokull volcanic plume in April2010 was modelled by adding a volcanic source in Iceland.2020

This source was roughly estimated to deliver a near real-timeassessment of the ash plume dispersion which was validatedagainst monitoring stations and lidar remote sensing (Coletteet al. (2011a)).

10.3 Long term evaluation2025

The long term evaluation of a model is probably the oldestway to estimate its accuracy during pollution events: not onlyshould the model be able to simulate specific and huge pol-lution events, it also has to calculate accurately low concen-trations in absence of pollution.2030

For a CTM, the air quality networks are able to continu-ously deliver hourly concentrations of O3, NO2 and particu-late matter over the past decades. Depending on the modeldomain and resolution, CHIMERE results have been com-pared to surface data. Enjoying the increase of computational2035

capabilities, these comparisons evolve from a few surfacedataset to complete hourly validation over several years. Thefirst comparisons were done for a few days and over a lim-ited region, the Paris area, Menut et al. (2000b) to a full year(Hodzic et al. (2004), Hodzic et al. (2005)), then a few month2040

and over larger domains: for example, Spain in Vivanco et al.(2008) and the western Europe in Colette et al. (2011b) andWilson et al. (2012). In Europe, the AirBase network is usedin many studies to compare CHIMERE results to surface ob-servations of O3, NO2, SO2, PM10, PM2.5. To extend the2045

comparison to vertical profiles, several data types were used:for ozone, vertical profiles of sondes and ozone analyseraboard commercial aircraft (MOZAIC/IAGOS) were com-pared to model outputs, Coman et al. (2012), Zyryanov et al.(2012).2050

Satellite data have also been used for long-term evaluationstudies as listed in Table 9 . CHIMERE has been comparedto NO2 satellite observations (tropospheric columns) fromSCIAMACHY (Blond et al. (2007)), OMI (Huijnen et al.(2010)) and GOME Konovalov et al. (2005). CHIMERE2055

has been evaluated against AOD measurements from MODISand POLDER satellites (Hodzic et al. (2006c), Hodzic et al.(2007)).

Satellite ozone observations from satellite and AQ mod-els can now be used synergistically either to evaluate mod-2060

els, to interpret satellite observations or to constrain modelsthrough assimilation. In this way, Konovalov et al. (2006)used SCIAMACHY NO2 columns to optimise NOx surfaceemissions. Zyryanov et al. (2012) have evaluated CHIMEREand MACC AQ models against IASI 0-6km ozone columns2065


Project Main goals (+references) Model improvementESQUIF∗ Regional photochemistry (Paris area, France) - Menut et al.

(2000b), Vautard et al. (2003), Hodzic et al. (2006c)Regional modelling of aerosol and gaseouspollutants, and aerosol optical properties

ESCOMPTE∗ Regional photochemistry (Marseille area, France) - Menut et al.(2005a)

Regional cartesian mesh model, for gaseouspollutants only

GEMS,MACC

Monitoring Atmospheric composition and climate - Hollingsworthet al. (2008)

European cartesian mesh model

AMMA∗ Mineral dust (western Africa) - Menut et al. (2009a) Mineral dust additionMILAGRO∗ Regional pollution (Mexico City) - Hodzic et al. (2010b), Hodzic

and Jimenez (2011)Secondary Organic Aerosols

CIRCE Climate change and impact research: The Mediterranean Environ-ment - Curci et al. (2009), Bessagnet et al. (2008)

Feedbacks between ozone and vegetation. Im-plementation of MEGAN.

GEOMON Global Earth Observation and Monitoring Pollution trends over EuropeMEGAPOLI∗ Regional pollution (Paris area) - Royer et al. (2011) Fast chemistry and aerosols over the Paris areaEC4MACS European Consortium for Modelling Air Pollution and Climate

Strategies - www.ec4macs.euEmissions reduction scenarios

AQMEII Models intercomparisons over Europe and United-States - Pirovanoet al. (2012), Solazzo et al. (2012b)

US domain

CITYZEN Impact of megacities on air pollution, trends analysis Colette et al.(2011b)

Emission mapping

ATOPICA Atopic diseases in changing climate, land use and air quality Pollens addition in the model

Table 8. International projects involving CHIMERE (in chronological order). Names with a ∗ correspond to field campaigns for whichmeteorological and chemical data were used to validate CHIMERE.

Satellite Parameter GoalSCIAMACHY,OMI, GOME

NO2, HCHO Improve total column and bio-genic emissions

MODIS,POLDER

Surfaceproperties

Fires emissions calculation.Improve aerosol emissionsand transport

IASI O3, CO Improve the troposphericcolumns and vertical distribu-tion

CALIOP Dust,aerosols

Improve the vertical transport

Table 9. Satellite data used for model/data comparisons and anal-ysis.

over one summer. Coman et al. (2012) have recently shownthe possibility to use IASI observations to correct CHIMEREmodel using an assimilation approach. Dufour et al. (2009)and Curci et al. (2010) applied inverse modelling of foma-ldehyde columns(SCIAMACHY and OMI, respectively) to2070

estimate and validate biogenic VOC emissions at the Euro-pean scale.

10.4 Models inter-comparisons and ensembles

CHIMERE has been involved and tested in a number of inter-comparison studies with other air pollution models. In order2075

to evaluate the sensitivity of air quality to emission controlscenarios and their uncertainty, several models were evalu-ated over a reference year and then used with emission sce-

narios. This work was conducted in the framework of theClean Air For Europe program (Cuvelier et al. (2007); Thu-2080

nis et al. (2007)), over four European cities. The evaluationpart of the project (Vautard et al. (2007)) showed the largespread of model simulations for ozone close to the sourcesdue to combined effects of titration and poor representa-tion of lower-layer mixing in stable boundary layers. This2085

spread gave rise to a spread of response to emission con-trol scenarios (Thunis et al. (2007)). At continental scale,this spread was less marked (Van Loon et al. (2007)). In thisinter-comparison over Europe, CHIMERE was found to haveamong the best skills for ozone daily maxima but to over-2090

estimate nighttime ozone concentration, leading to a generalpositive bias, not generally shared by other models. This biasis thought to be due in large parts to overestimation of mix-ing in stable conditions. For particulate matter, CHIMEREwas shown to exhibit a negative bias, shared by other mod-2095

els, at least over several high wintertime episodes in Ger-many (Stern et al. (2008)). Such biases were also found in amore recent and extensive inter-comparison over two conti-nents and a full evaluation year (Rao et al. (2011); Solazzoet al. (2012b), Solazzo et al. (2012a)). In this unprecedented2100

exercise, models used in Europe and in North America wereconsidered. CHIMERE also participated to the first multi-model decadal air quality assessment in the CityZen project(Colette et al. (2011b) and proved to be in-line with otherstate-of-the-art tools. The same biases as previously reported2105

(positive for mean ozone and negative for particulate mat-ter) where obtained by the majority of the models. As faras CHIMERE was concerned, the usual strength in capturing


ozone variability through a good temporal correlation wasfound.2110

In these inter-comparison studies, the potential to use theseensemble of models to (i) improve the simulation of air pol-lutant concentrations by a proper averaging of results and (ii)to estimate the uncertainty in the simulations was evaluated(Solazzo et al. (2012b)). It was shown in particular that the2115

average of model results outperformed each individual modelresult (Van Loon et al. (2007)), and that for most pollutantsthe spread of models was representative of their uncertaintyand skill (Vautard et al. (2009)).

11 Data assimilation2120

The forward chemistry-transport model uses meteorologyand emissions as forcings to calculate pollutant concentra-tions fields. The hybridation consists in assimilating obser-vations during a simulation in order to: (i) optimise one ofthe forcing parameter (inverse modelling), (ii) build more re-2125

alistic concentrations fields (data assimilation analysis).

11.1 Sensitivity studies

For several of the mentionned applications, the adjoint modelis a powerful tool and may be used for sensitivity studies andinverse modelling. The first adjoint of CHIMERE was de-2130

veloped in 1998 and used to optimise the boundary condi-tions of the box-model version, Vautard et al. (2000), Menutet al. (2000a). An updated version was developed whenCHIMERE was modified to use a cartesian mesh, Menut(2003). CHIMERE being updated every year, the adjoint2135

model should follow the same evolution: unfortunately, thelast adjoint version was the one developed for the gaseousspecies. When the aerosols were added, Bessagnet et al.(2004), the adjoint part was not upgraded. Currently, a newbranch of the model’s adjoint is under development: the ver-2140

sion applied to regional CO2 fluxes inversion over westernEurope Broquet et al. (2011) is being parallelised.

The adjoint model was used to quantify to which inputparameter the modelled concentrations are sensitive, Menut(2003). The studies done with CHIMERE were for pollution2145

in the Paris area: the calculations were performed to calcu-late the sensitivity of O3, Ox and NOx to various meteoro-logical parameters and surface emissions fluxes (per activ-ity sectors). An example of synthesized results is presentedin Figure 14 . For one specific concentration (here NOx at2150

09:00 UTC and O3 at 15:00 UTC in the Paris center gridcell), the adjoint model enables to calculate the sensitivity tothe whole domain emissions of traffic and solvents is calcu-lated. This shows that NOx is more sensitive to emissionsthan O3, quantifying the direct effect of modelling primary2155

and secondary species. The sensitivity may be positive (a di-rect addition of NOx by traffic emissions will increase instan-taneously NO concentrations) or negative (O3 titration by

Fig. 14. Time series of the sensitivity of pollutants (NOx and O3) toanthropogenic surface emissions (from ’traffic’ and ’solvents’ activ-ity sectors).

NO2). This kind of study allowed us to classify the most im-portant parameters in a chemistry-transport model, depend-2160

ing on the modelled pollutants, the location and the time.

Another way to make sensitivity studies is Monte Carlomodelling. The same pollution events were studied over theParis area. The results allowed us to quantify the variabilityof pollutants and thus to refine the uncertainty of the mod-2165

elled concentrations as a function of the uncertainties of theinput parameters, Deguillaume et al. (2008).

11.2 Inverse modelling of emission fluxes

Over the Paris area, the inversion of anthropogenic emissionswas done for specific pollution events and seasonal simula-2170

tions. These studies were an opportunity to develop a newapproach. The surface measurements used are less numerousthan the grid points to invert: this weak-constrained prob-lem is often circumvented at the global scale by inverting theemissions of the same species as the measured one and con-2175

sidering large areas and long timescale. For regional studiesand photo-oxidant pollution, the time and spatial variabil-ity is large. A methodology of dynamical areas was there-fore developed and applied, Pison et al. (2006), Pison et al.(2007). For the Paris area, the results enabled us to opti-2180

mise the diurnal profiles of the emissions and show that thecity center emissions were over-estimated in emission inven-tories whereas the suburban emissions were often underesti-mated. An example of optimised coefficients is displayed inFigure 15 for NO emissions at 05:00 UTC, when the traffic2185

becomes important.


Fig. 15. Inverse modeling of the NO emissions in the Paris area forAugust, 7 1998 (ESQUIF campaign). The results are multiplicativefactors to apply to the emissions flux hour by hour. A value of 1means there is no change to apply to the fluxes, when, for example,a value of 0.5 means it is needed to divide by two the NO emissionsfluxes to reduce the differences between the measured and modeledsurface concentrations of NOx.

11.3 Analysis of concentration fields

Following the approach developed in meteorology andoceanography, data assimilation has been applied to air qual-ity since the beginning of this century. In the case of ozone,2190

ground based observations from air quality networks havebeen used to correct regional CTMs (Hanea et al. (2004);Wu et al. (2008)). In the case of the CHIMERE model,Blond et al. (2003) and Blond and Vautard (2004) have de-veloped an optimal interpolation method where they used an2195

anisotropic statistical interpolation approach to determine aclimatological background covariance matrix (Blond et al.(2003)). This matrix allows them to give weights to inno-vations (differences between model and observations) and topropagate this information where no observations are directly2200

available. Over a European domain they improve RMSE byabout 30%. As already stated by Elbern and Schmidt (2001),forecasts using these analyses as initial conditions were onlyslightly improved in few particular cases. The work ofBlond and Vautard (2004) has been implemented in the PRE-2205

VAIR platform that produces operationally ozone analysis(Honore et al. (2008)). Such analyses are also produced inthe framework of the FP7/MACC-II project as well as PM10

analysis derived from a similar approach (http://www.gmes-atmosphere.eu/services/raq/).2210

Since 2006, an ensemble Kalman filter (EnKF; Evensen(1994)) has been coupled to the CHIMERE model. It is alsoa sequential assimilation method but it allows calculating atime-evolutive background covariance matrix that takes intoaccount the variability of model errors with time and space.2215

It is based on a Monte Carlo approach using an ensemble of

forward simulations to calculate the background covariancematrix. To do so, we have followed the precursor work ofHanea et al. (2004) that coupled an EnKF to the LOTOS-EUROS model. One of our goals was to assimilate satellite2220

data that would complement surface observations. Indeed,since 2006, the IASI instrument on board the METOP plat-form (Clerbaux et al. (2009)) allows us to observe ozone con-centrations in the lower atmosphere (0-6 km partial columns)with a good accuracy (Eremenko et al. (2008); Dufour et al.2225

(2012)). Coman et al. (2012) have shown that assimilatingIASI 0-6km columns in the CHIMERE model allows cor-recting significantly tropospheric ozone fields, Figure 16 .Corrections were higher at about 3-4 kilometers height wherethe instrument and retrieval method exhibit maximum sensi-2230

tivity. In spite of a reduced sensitivity of the instrument in theplanetary boundary layer, Coman et al. (2012) also showedthat surface ozone fields were systematically improved.

Fig. 16. Assimilation with CHIMERE of ozone profiles recordedby the IASI instrument and comparisons of the obtained gain withMOZAIC data (Data courtesy of A.Coman).

In the near future, it is planned to produce analyses by as-similating simultaneously surface and satellite measurements2235

of ozone. Such a product could be of great interest to studytropospheric ozone variability and trends especially in re-gion were in situ observations are scarce such as the Mediter-ranean basin. This CHIMERE-EnKF software will be testedoperationally during the FP7/MACC-II project. Moreover,2240

assimilation of other species such as NO2 and CO (fromsatellite) is planned in the framework of the same MACC-II project.

12 Forecast

Air quality forecast is one main goal of chemistry-transport2245

modelling, but it is also a specific way to improve and to vali-


date models (Menut and Bessagnet (2010)). Indeed, forecastmakes it possible to quantify day by day the model accuracyand to measure its sensitivity to the various parameters andparameterisations.2250

12.1 Experimental forecast

Fig. 17. Time series of dust forecast for the period of March, 17 to23 2006. The results are for dust surface concentations in Rome,Italy.

Since its early developments, CHIMERE has been usedboth for analysis and forecast. For the release of every newmodel version, the development priorities were driven by theresults of test case analyses and daily experimental forecasts.2255

A recent example of this is the COSY project which aims atproducing systematic comparisons between observations anda set of CHIMERE forecasts and making them available ona web site (www.lmd.polytechnique.fr/cosy/). This projecthas provided sensitivity analyses of modelled concentration2260

fields when using the different land surface schemes of theWRF meteorological driver, Khvorostyanov et al. (2010):this leads to a better understanding of the impact of LandSurface Models (LSM) on modelled surface concentrationsand thus quantify the concentration biases only due to mete-2265

orological surface fluxes. Another example comes from dustmodelling. Figure 17 presents modelled dust surface con-centrations over Rome, Italy, for several leads of the sameperiod (Menut et al. (2009a)). The variability of the meteo-rological forecast impacts directly the emissions and, finally,2270

the remote surface concentrations: the daily forecasted max-ima may show differences up to a factor of 2. Unfortunately,this variability is of the same order of magnitude as the back-ground particle concentrations often recorded in Europe: thisclearly shows that the forecast of mineral dust transport over2275

Europe remains a challenging scientific problem.

12.2 Operational forecast

CHIMERE is implemented on several air quality platformswhich provide daily forecasts up to 3 days ahead for a setof regulatory pollutants (e.g. O3, NO2, PM10, PM2.5).2280

Firstly developed as experimental platforms, such tools(www.prevair.org; Rouıl et al. (2009), Menut and Bessagnet(2010)) became operational in France after the 2003 summerheat wave (Vautard et al. (2005)) and are foreseen to be op-erational in Europe at the end of the FP7 project MACCII in2285

2014. Forecasts of pollution above accepted threshold levelsare essential for public information.

Moreover, air quality models offer two essential function-alities. They can help to identify the reasons why pollutantconcentrations increase (giving for instance PM speciation),2290

and the results of operational runs conducted with emissioncontrol scenarios allow selecting the most efficient measures.As a consequence, these modelling tools provide support tonational authorities on air quality management, and they as-sist the selection of regulatory measures that may be efficient2295

to limit the intensity of pollution episodes. In this context,CHIMERE is now involved in the prototype toolbox ded-icated to air quality episode management in the MACC-IIproject.

Another contribution of such platforms to air quality issues2300

is the provision of daily assessments of the model ability topredict pollutant concentrations. Daily scores are indeed animportant parameter, which gives insights into the researchefforts that need to be made to improve the model behaviour.Current research efforts are thus focusing on the implementa-2305

tion of dynamical input data such as biomass burning emis-sions (Kaiser et al. (2012)) or emissions from agriculturalactivities Hamaoui-Laguel et al. (2012)). In both cases, thestarting date of emissions is crucial for forecasting their im-pacts in time, place and magnitude on pollutant concentra-2310

tions. The wintertime PM episodes forecasts are also im-proved by taken into account the variability of climatic con-ditions when calculating the wood burning emissions fromresidential heating.

Finally, in the MACC-II project, it was shown that im-2315

port of pollutants due to cross-Atlantic ozone plumes orAfrican dust plumes could contribute significantly to the Eu-ropean pollutant levels computed with CHIMERE (Menutet al. (2009b)), and that the implementation in CHIMERE ofnear real time boundary conditions delivered by global mod-2320

els provided reliable pollutant background concentrations tothe European regional domains. Currently, CHIMERE fore-casts are used by more than 50 local agencies in Europe, ei-ther to refine the PREV’AIR forecasts using their own mod-elling platform or to produce a local air quality index value2325

(such as the CITEAIR index, http://www.airqualitynow.eu).


13 Specific applications

13.1 Chemistry in fire plumes

Quantifying the impact of fires on air quality requires notonly accurate emissions but also a realistic representation of2330

the chemical processes occuring inside the plumes.Fire plumes, besides having high spatial variability, con-

sist of large amounts of trace gases (including the main ozoneprecursors) and aerosols.

The corresponding emissions are included in the emission2335

inventory. However these dense plumes will induce a sig-nificant modification of the UV light reaching the surface,and thereby the photolysis rates for photochemical reactions.For example, Alvarado and Prinn (2009) estimate that, if thiseffect is accounted for, the ozone levels in plumes from sa-2340

vanna fires in South Africa decrease by 10-20%, and Hodzicet al. (2007) estimate a 10-30% decrease for Portuguese for-est fires.

CHIMERE currently uses tabulated photolysis rates pre-calculated using the Troposphere Ultraviolet and Visible2345

(TUV) model (Madronich et al. (1998) for different cloudcover situations. In order to account for the impact of denseplumes (critical for fires but also dust and anthropogenicsources), the photolysis rates need to be reevaluated depend-ing on the evolution of the aerosol optical depth (AOD) at2350

different vertical levels and locations. Several approacheshave been used for CHIMERE. A simplified parameterisa-tion based on satellite observations of the AOD has beenused by Hodzic et al. (2007), and Konovalov et al. (2011)have refined this approach by recalculating the photolysis2355

rates online with the TUV model using the observed AODas constrain. More recently, an online calculation of theaerosol properties has been implemented and coupled to theTUV model (Pere et al. (2011)). It is currently being inte-grated in the standard version of CHIMERE, with a numeri-2360

cal optimisation of the calculation of aerosol optical proper-ties to maintain computational efficiency. Large variations inozone production are also observed depending on PAN for-mation, heterogeneous chemistry or oxygenated volatile or-ganic compounds amounts for different fire situations (Jaffe2365

and Wigder (2012), Konovalov et al. (2012)), that are still notwell represented by CTMs. Once the effect of aerosols on ra-diation is included, further analysis will be undertaken on thechemical evolution in fire plumes simulated by CHIMERE.

13.2 Heavy metals2370

Metals are considered as important pollutants that can beresponsible for a range of human health effects. Diseasessuch as cancer, neurotoxicity, immunotoxicity, cardiotoxic-ity, reproductive toxicity, teratogenesis and genotoxicity canbe related with the presence of metal particles in the air (HEI2375

(1998); EPA (1999)). Organisms can assimilate the particlesvia inhalation (because particles settle into bronchial regions

Fig. 18. Mean annual concentrations of lead (µg m−3) for 2009over Spain and Portugal.

of the lungs) or ingestion (because the particles are depositedand accumulated on soils or water, and this accumulationproduces an increase of the risk of future exposure through2380

food). In Europe, Directive 2008/50/CE sets an annual limitvalue of 500 ng/m3 for Pb. Annual target levels for As, Cdand Ni are regulated by Directive 2004/107/CE (6 ng/m3 forAs, 5 ng/m3 for Cd, 20 ng/m3 for Ni). For other metals (withthe exception of mercury), no normative is available.2385

In the atmosphere, metals are attached to particles, es-pecially those in the fine fraction (Milford and Davidson(1985); Allen et al. (2001); Molnar et al. (1995); Kulogluand Tuncel (2005)). A preliminary description of heavy met-als (Pb, Cd, As, Ni, Cu, Zn, Cr and Se) air concentration has2390

been implemented in a dedicated version of the CHIMEREmodel. At this stage, these metals are treated as inert fineparticles. According to Finlayson-Pitts and Pitts (2000), theaerodynamic mass median diameters for Pb, Cd, As, Ni, Cu,Zn, Cr and Se are 0.55, 0.84, 1.11, 0.98, 1.29,1.13, 1.11 and2395

4.39µm, respectively. As in this approach all the metals areconsidered as fine particles, more refinement could be neces-sary for Se. Physical processes such as anthropogenic emis-sions, transport, mixing and deposition are considered. Forsome of these metals, such as Pb and Cd, the inert status2400

consideration is generally adopted. They are believed to betransported in the atmosphere with no change in their chem-ical and aggregate state (Ryaboshapko et al. (1999)). Forother metals this approach must be reconsidered and reac-tions in the aqueous phase could be required. In the case of2405

Cr, Seigneur and Constantinou (1995) have highlighted theimportance of the reactions converting Cr(III) to Cr(VI) andvice-versa associated to particle and droplet chemistry. ForNi, aqueous phase chemistry could also be important. Re-garding arsenic, airborne particulate matter has been shown2410

to contain both inorganic and organic arsenic compounds


(Johnson and Braman (1975); Attrep and Anirudhan (1977)).A preliminary study applying this heavy metal CHIMERE

version has been presented in Vivanco et al. (2011), wherethe model performance was evaluated for Spain. Figure 182415

presents the mean annual concentrations of lead for 2009.This preliminary version has also recently been applied for aEuropean domain at a horizontal resolution of 0.2o for 2008,Gonzalez et al. (2012). Important limitations were set formetal emissions. Only anthropogenic sources have been con-2420

sidered, although some metals can be released into the envi-ronment by both natural sources and human activities. More-over, only Cd and Pb emissions are available in the EMEP ex-pert database for 2008 (Vestreng et al. (2009)); for the othermetals, emissions were taken from TNO totals estimated for2425

2000, Van der Gon et al. (2005), except in the Spanish partof the domain, where emissions were provided by the Span-ish Ministry of Environment and Rural and Marine Affairsfor 2006. Original emissions were spatially and temporallydisaggregated and adapted to the simulated domain by taking2430

into account the land use information of GLCF (Global LandCover Facility, http://change.gsfc.nasa.gov/create.html). Inthe case of TNO emissions, only the total amount for eachEMEP grid-cell was available. The temporal disaggregationof these totals is normally performed in the CHIMERE emis-2435

sion processor considering the SNAP activity. As we didnot have this information for the metals coming from TNOdatabase, we used a SNAP disaggregation similar to PM2.5particles for those species. We also used a temporal profilesimilar to PM fine particles. A better knowledge of the tem-2440

poral behavior of metal emissions for each SNAP activitywould reduce the input errors. Other aspects, such as bound-ary conditions should also be considered in order to improvemodel results. At this stage, no boundary conditions were ap-plied, as no information on many of these metals is available2445

from global models.

13.3 Pollens

A new research direction involving CHIMERE is modellingthe dispersion of pollen grains in the atmosphere. The preva-lence of pollen allergy in European countries is estimated be-2450

tween 12 and more than 35% (Burney et al. (1996)). Thequantity of some highly allergenic pollens, such as ragweed,is increasing and seems to correlate with allergic diseases(Rybnicek and Jaeger (2001)). Conclusions from a numberof European and international climate research projects sug-2455

gest that future climate change and variability may stronglyaffect pollen emission and dispersal in Europe (Christensenet al. (2007)), thereby influencing the prevalence of atopicdiseases.

The European (FP7) project Atopica aims to evaluate, us-2460

ing numerical modelling, statistical data analysis, and labo-ratory experiments, the influence of changes in climate, airquality, land use, and the subsequent distribution of invasiveallergenic plant species and allergic pollen distribution on hu-

Fig. 19. Birch pollen concentrations (grains m−3) simulated withCHIMERE for April 12, 2007.

man health. Modelling of ragweed and birch pollen emission2465

and dispersion in the atmosphere using CHIMERE plays akey role in this project.

Pollen grains are about 5-50 times larger in size than con-ventional atmospheric aerosols. The scale analysis (Sofievet al. (2006)) shows that the assumption of pollen grains be-2470

ing transported together with air masses following the air-flow, including small turbulent eddies, still applies. Pollensare implemented in CHIMERE as a special aerosol type hav-ing a single size distribution bin of 20-22 µm depending onthe pollen type. The density is prescribed for a particular2475

pollen species and varies between 800 and 1050 kg m−3,which yields the sedimentation velocity of 1.2-1.3 cm s−1.Gravitational settling is the main deposition process for pol-lens and the only one considered in the model. The simulatedpollen grains are transported by the atmospheric circulation2480

and turbulent mixing, settled under the gravity, and washedout by rains and clouds, following the parameterisations al-ready implemented for other CHIMERE aerosols.

The main challenge for the state-of-the art pollen disper-sion modelling is accurate description of pollen emissions.2485

This requires a fair knowledge of plant distribution, phe-nology, and adequate assumptions regarding the sensitivityto meteorological factors, such as humidity, temperature,wind and turbulence. Unlike industrial pollutants or min-eral aerosols, pollen emissions depend not only on the in-2490

stantaneous meteorological conditions but also on the con-ditions during the pollen maturation within the plants beforethe pollination starts. Additional factors such as the CO2 andO3 concentrations can influence pollen production and emis-sions Rogers et al. (2006), Darbah et al. (2008).2495

Pollen episodes have been simulated with CHIMERE forragweed (Chaxel et al. (2012)) and for birch, the latter us-ing the emission methodology developed by Sofiev et al.


(2013). Figure 19 shows daily mean birch pollen concen-tration simulated with CHIMERE for April 12, 2007, during2500

the seasonal concentration maximum in Paris. CHIMEREwas forced by the WRF model in the forecast mode with-out nudging at 15 km resolution for the North of France.The model concentration in Paris peaks at 390 grains m−3

on April 12, while the RNSA observations show the peak2505

of 965 grains m−3 on April 14. The model underestimationcan be related to the birch distribution map used for pollenemission calculation: its French contribution is mostly basedon the satellite data prone to large uncertainties Sofiev et al.(2006).2510

13.4 Subgrid scale exposure modelling for health im-pact assessment

Air quality models, by integrating different emission sce-narios, are valuable tools for the evaluation of alternativemitigation policies and possible adaptation strategies with2515

respect to public health and climate change. CHIMEREmodel is involved in two different projects on the evalua-tion of the health impact of air quality under changing condi-tions (ACHIA and ACCEPTED). This ongoing research re-quires several model developments in order to account for2520

the variability of exposure at the intra-urban scale. A sub-grid scale module has been added to the classical computa-tion of the grid-averaged pollutant concentration based on theReynolds-average approach. Since grid cell surfaces covergenerally a few square kilometers in meso-scale models, the2525

large heterogeneities in emissions over urban areas can not berepresented, although this may result in concentrations nearsources that are very different from the grid averaged levels.From a health outcome perspective, it is important to knowhow much local concentrations deviate from grid-averaged2530

values due to the proximity to emission sources (e.g. nearroads, in residential zones, industrial parks etc.).

To address the issue of subgrid-scale emission hetero-geneity, an emission scheme has been developed and imple-mented in the CHIMERE model (Valari and Menut (2010)).2535

Instead of adding together emissions from all sources and allactivity sectors at each model grid cell we split them intofour different categories: traffic, residential, outdoor activ-ities (entertainment) and all other sources (including pointsources). Grid area fractions corresponding to each type of2540

emission are calculated based on high resolution landuse data(CORINE land cover at 100m resolution). Thus, instead ofa single emission from each model grid cell

∑Ei/A, four

different emission scenarios A×Ei/Ai are used. At eachmodel time-step, and for the grid cells where we are inter-2545

ested in applying the subgrid scale calculation, instead ofthe single ’grid-averaged’ scenario, concentrations are cal-culated following all four subgrid scale scenarios. At the endof the time-step, weighted averages of the four estimates forall model species are propagated to the next time-step.2550

Fig. 20. Surface concentrations of NO2 time series, using the sub-grid scale variability module. The symbols represent AIRPARIFmeasurements of two stations, one background and one traffic, lo-cated in the same CHIMERE cell in Paris, with a horizontal res-olution of 3kmx3km. The plain lines represent the correspondingmodelled surface concentrations and show the subgrid scheme isable to reproduce the large variability of NO2 depending on thesources in an urban environment. After Valari and Menut (2010).

In Figure 20 , modelled NO2 concentrations are com-pared to surface measurements of the Ile-de-France air-quality network AIRPARIF. Model resolution is 3kmx3kmand results are shown for a model grid cell with in whichlay two AIRPARIF sites: a traffic monitor (red circles) and2555

a background monitor (blue crosses). The black line corre-sponds to the standard model grid-averaged concentrationswhereas colored lines stand for the results of the subgridscale scheme (red for the traffic sector and blue for the res-idential sector). The added value of the subgrid implemen-2560

tation is that instead of a single grid-averaged concentrationwe now have two additional estimates for what pollutant lev-els are near a road or inside a residential block inside the gridarea. This information is especially useful when air-qualitymodel outputs are used to estimate population exposure lev-2565

els. Subgrid model concentrations weighted with activitydata on the time people spend at home, at the office or intransit, give an estimate of personal exposure. This methodhas been applied to the Paris area in Valari et al. (2011).

14 Conclusions and perspectives2570

This paper presented the complete schemes and parame-terizations implemented in CHIMERE (v.2013), a regionalchemistry-transport model dedicated to atmospheric compo-sition studies. The model is continuously under developmentand this article represents the reference model description.2575


CHIMERE was initially developed for regional and short-term (a few days) pollution episodes. During the last years,new projects and demand gave rise to several needs for modeldevelopments, leading to changes in (i) the spatial domainresolution and size, (ii) the simulations duration and (iii) the2580

processes and parameters covered. Spatially, the model do-mains were extended to near semi-hemispheric scales forconcentration plumes due to mineral dust, forest fires andvolcanic emissions. At the same time, urban versions weredeveloped to better understand and quantify the impact of air2585

pollution on health.As perspectives, the model will continue to evolve by re-

fining the schemes and parameterizations used. In addition,the next version will be an on-line version. Indeed, his-torically, and for computational reasons, many CTMs such2590

as CHIMERE have been essentially used offline, i.e. theyare forced with precalculated meteorological fields and sur-face state. In order to consider the numerous feedbacks be-tween atmospheric, radiative and chemical processes, the de-velopment of platforms coupling online meteorology, chem-2595

istry and vegetation has been identified as a priority for bothresearch and forecast applications. One next big step inCHIMERE development will be to build an online-coupledplatform putting together a meteorological model (WRF),a ground-vegetation-hydrology model (ORCHIDEE) and a2600

CTM (CHIMERE). This coupling will be done through theOASIS coupler, which will offer greater flexibility and com-putational advantages compared to a direct hard-coded im-plementation of a supermodel that includes all these threemodels. The resulting modelling platform will be used for2605

fundamental research purposes, for assessing the importanceof the feedback between meteorology, vegetation and chem-istry, and it will be available for research and forecast appli-cations.

Acknowledgments:2610

The CHIMERE development team thanks Mireille Lat-tuati, Cecile Honore, Hauke Schmidt, Claude Derognat,Laurence Rouıl, among others, for their contribution informer CHIMERE developments. We also acknowledgeEric Chaxel for the WRF model meteorological inter-2615

face, Christian Seigneur and Betty Pun (AER) for the re-duced SOA scheme, Mian Chin and Paul Ginoux (NASA)for the GOCART aerosols concentrations fields used asboundary conditions, Sophie Szopa and Didier Hauglus-taine (IPSL/LSCE) for the LMDz-INCA gas concentrations2620

elds used as boundary conditions, Athanasios Nenes and theISORROPIA team for the free use of ISORROPIA model,the EMEP-MSCWEST team (http://www.emep.int) for an-thropogenic emissions database, Cathy Liousse, Bruno Guil-laume and Robert Rosset (LA / OMP, http://www.aero.obs-2625

mip.fr) for organic and black carbon emissions, the EMEP-MSCEAST team for benzo(a)pyrene, benzo(k)βuoranthene

and benzo(b)βuoranthene data (http://www.msceast.org).The developpers also thank all the users with whom theyhave regularly interesting discussions.2630

Appendix A MELCHIOR2 Chemical Mechanism

This appendix presents the MELCHIOR2 gas-phase chem-ical mechanism and all related model species used inCHIMERE. In Table A1 , the notes correspond to:

• a Hydrocarbon model species (except for CH4 and C2H4)2635

represent groups of hydrocarbons with similar reactivity.As ”definition”, a typical representative of this group isgiven. This is also true for the degradation products.• b For the concept of ”chemical operators” see Carter

(1990) and Aumont et al. (2003).2640


Table A1. Species list of the reduced MELCHIOR2 gas-phasechemical mechanism.

Symbol Full name

Inorganic compoundsO3 ozoneH2O2 hydrogen peroxideOH hydroxy radicalHO2 hydroperoxy radicalNO nitrogen oxideNO2 nitrogen dioxideNO3 nitrogen trioxideN2O5 dinitrogen pentoxideHONO nitrous acidHNO3 nitric acidCO carbon monoxideSO2 sulphur dioxide

Hydrocarbon speciesa

CH4 methaneC2H6 ethaneNC4H10 n-butaneC2H4 etheneC3H6 propeneOXYL o-xyleneC5H8 isopreneAPINEN α-pineneBPINEN β-pineneLIMONE limoneneTERPEN terpenesHUMULE HumuleneOCIMEN Ocimene

CarbonylsHCHO formaldehydeCH3CHO acetaldehydeCH3COE methyl ethyl ketoneGLYOX glyoxalMGLYOX methyl glyoxalCH3COY dimethyl glyoxalMEMALD unsaturated dicarbonyls,

reacting like 4-oxo-2-pentenalMVK methyl vinyl ketoneMAC methacrolein

Table A1. Species list of the reduced MELCHIOR2 gas-phasechemical mechanism (continued)

Symbol Full name

Organic nitratesPAN peroxyacetyl nitrateCARNIT nitrate carbonyl taken as α-nitrooxy acetoneISNI unsaturated nitrate from isoprene degrada-

tion

Organic peroxidesCH3O2H methyl hydroperoxidePPA peroxy acetyl acid

Peroxy radicalsCH3O2 methyl peroxy radicalCH3COO peroxy acetyl radical

Operatorsb

oRO2 representing peroxy radicals from OH at-tack to C2H6, NC4H10, C2H4, C3H6,OXYL, CH3COE, MEMALD, and MVK

oROOH representing organic peroxides fromoRO2+HO2 reactions

obio representing peroxy radicals produced byC5H8 and APINEN + OH reaction

obioH representing biogenic organic peroxidesfrom obio+HO2 and obio+obio reactions

oPAN representing PAN homologue compounds(except PAN)

PANH representing results from oPAN+HO2 reac-tion

toPAN representing results from oPAN+NO2 reac-tion

oRN1 representing organic nitrate peroxy radicalsfrom NO3 attack to C2H4, C3H6, C5H8,APINEN, BPINEN, LIMONE, TERPEN,OCIMEN, HUMULE and OH attack toISNI


Table A2. Reaction list of the reduced MELCHIOR2 chemical mechanism

Reactions Kinetic constants Ref.O3+NO→ NO2 Ae−B/T ,A=1.8e-12,B=1370 [1]

O3+NO2→ NO3 Ae−B/T ,A=1.2e-13,B=2450 [1]

O3+OH→ HO2 Ae−B/T ,A=1.9e-12,B=1000 [1]

O3+HO2→ OH Ae−B/T ,A=1.4e-14,B=600 [1]

NO+HO2→ OH+NO2 Ae−B/T ,A=3.7e-12,B=-240 [1]NO2+OH+M→ HNO3 troe(3.4e-30,0,3.2,4.77e-11,0,1.4,0.30) [2],n9HO2+OH→ H2O Ae−B/T ,A=4.8e-11,B=-250 [1]

H2O2+OH→ HO2 Ae−B/T ,A=2.9e-12,B=160 [1]

HNO3+OH→ NO3 Ae−B/T ,A=5.5e-15,B=-985 [3], n1CO+OH→ HO2+CO2 Ae−B/T (300/T)N , A=2e-13, B=0, N=1 [3], n1HO2+HO2→ H2O2 Ae−B/T ,A=2.2e-13,B=-740 [1], n1HO2+HO2+H2O→ H2O2 Ae−B/T ,A=4.52e-34,B=-2827 [1], n1NO3+HO2→ NO2+OH 4e-12 [1]NO3+H2O2→ HNO3+HO2 2e-15 [3]

NO3+NO→ 2*NO2 Ae−B/T ,A=1.8e-11,B=-110 [1]

NO2+NO3→ NO+NO2 Ae−B/T ,A=4.5e-14,B=1260 [3]NO2+NO3+M→ N2O5 troe(2.7e-30,0,3.4,2e-12,0,-0.2,0.33) [1],n9N2O5+M→ NO3+NO2 troe(1e-3,11000,3.5,9.7e14,11080,-0.1,0.33) [1],n9N2O5+H2O→ 2*HNO3 2.6e-22 [4]N2O5+H2O+H2O→ 2*HNO3 2e-39 [4]NO+OH+M→ HONO troe(7.e-31,0,2.6,1.5e-11,0,0.5,0.6) [1],n9HONO+OH→ NO2 Ae−B/T ,A=1.8e-11,B=390 [1]

NO+NO+O2→ 2*NO2 Ae−B/T ,A=3.30E-39,B=-530.0 [1]NO2→ HONO 0.5*depo(NO2) [18]

SO2+CH3O2→ H2SO4+HCHO+HO2 4e-17 [3]SO2+OH+M→ H2SO4+HO2 troe(4e-31,0,3.3,2e-12,0,0,0.45) [3],n9CH4+OH→ CH3O2 Ae−B/T ,A=2.3e-12,B=1765 [1]

C2H6+OH→ CH3CHO+oRO2 Ae−B/T ,A=7.9e-12,B=1030 [1]

NC4H10+OH→ 0.9*CH3COE +0.1*CH3CHO +0.1*CH3COO+0.9*oRO2

Ae−B/T (300/T)N ,A=1.36e-12, B=-190, N=-2 [5]

C2H4+OH+M→ 2*HCHO+oRO2 troe(7e-29,0,3.1,9e-12,0,0,0.7) [1],n9C3H6+OH+M→ HCHO+CH3CHO+oRO2 troe(8e-27,0,3.5,3e-11,0,0,0.5) [1],n9OXYL+OH→MEMALD+MGLYOX+oRO2 1.37e-11 [5]

C5H8+OH→ 0.32*MAC +0.42*MVK +0.74*HCHO +obio Ae−B/T , A=2.55e-11, B=-410 [1]

APINEN+OH→ 0.8*CH3CHO+0.8*CH3COE+obio Ae−B/T , A=1.21e-11, B=-444 [17]

BPINEN+OH→ 0.8*CH3CHO+0.8*CH3COE+obio Ae−B/T , A=1.21e-11, B=-444 [17]

LIMONE+OH→ 0.8*CH3CHO+0.8*CH3COE+obio Ae−B/T , A=1.21e-11, B=-444 [17]

HUMULE+OH→ 0.8*CH3CHO+0.8*CH3COE+obio Ae−B/T , A=1.21e-11, B=-444 [17]

OCIMEN+OH→ 0.8*CH3CHO+0.8*CH3COE+obio Ae−B/T , A=1.21e-11, B=-444 [17]

TERPEN+OH→ 0.8*CH3CHO+0.8*CH3COE+obio Ae−B/T , A=1.21e-11, B=-444 [17]

HCHO+OH→ CO+HO2 Ae−B/T ,A=8.6e-12,B=-20 [1]

CH3CHO+OH→CH3COO Ae−B/T ,A=5.6e-12,B=-310 [1]MEMALD+OH→GLYOX+MGLYOX+oRO2 5.6e-11 [1]

CH3COE+OH→CH3COY+oRO2 Ae−B/T (300/T)N , A=2.92e-13, B=-414, N=-2 [1]GLYOX+OH→ 2*CO+HO2 1.1e-11 [1]MGLYOX+OH→ CH3COO+CO 1.5e-11 [1]

MVK+OH → 0.266*MGLYOX +0.266*HCHO+0.684*CH3CHO +0.684*CH3COO +0.05*ISNI +0.95*oRO2

Ae−B/T ,A=4.1e-12,B=-453 [6]

MAC+OH→ 0.5*CH3COE+0.5*CO2+0.5*oPAN Ae−B/T ,A=1.86e-11,B=-175 [6]

CH3O2H+OH→ CH3O2 Ae−B/T ,A=1.9e-12,B=-190 [1]

PPA+OH→ CH3COO Ae−B/T ,A=1.9e-12,B=-190 n2CH3O2H+OH→ HCHO+OH Ae−B/T ,A=1.e-12,B=-190 [1]

oROOH+OH→ 0.8*OH+0.2*oRO2 Ae−B/T ,A=4.35e-12,B=-455 n2obioH+OH→ OH 8e-11 n2PANH+OH→ 0.2*oPAN 1.64e-11 n2CARNIT+OH→ CH3CHO+CO+NO2 k(T)=Ae−B/T ,A=5.6e-12,B=-310


Table A2. Reaction list of the reduced MELCHIOR2 chemical mechanism (continued)

Reactions Kinetic constants Ref.C2H4+NO3→ 0.5*CARNIT+HCHO+oRN1 2e-16 [1]C3H6+NO3 → 0.5*CARNIT +1.5*HCHO +0.5*CH3CHO+0.5*HO2 +oRN1

9.45e-15 [1]

APINEN+NO3→ CH3CHO+CH3COE+oRN1 Ae−B/T ,A=1.19e-12,B=490 [8]

BPINEN+NO3→ CH3CHO+CH3COE+oRN1 Ae−B/T ,A=1.19e-12,B=490 [8]

LIMONE+NO3→ CH3CHO+CH3COE+oRN1 Ae−B/T ,A=1.19e-12,B=490 [8]

OCIMEN+NO3→ CH3CHO+CH3COE+oRN1 Ae−B/T ,A=1.19e-12,B=490 [8]

HUMULE+NO3→ CH3CHO+CH3COE+oRN1 Ae−B/T ,A=1.19e-12,B=490 [8]

TERPEN+NO3→ CH3CHO+CH3COE+oRN1 Ae−B/T ,A=1.19e-12,B=490 [8]C5H8+NO3 → 0.85*ISNI +0.1*MAC +0.05*MVK+0.15*HCHO +0.8*HO2 +oRN1

7.8e-13 [6]

HCHO+NO3→ CO+HNO3+HO2 5.8e-16 [1]CH3CHO+NO3→ CH3COO+HNO3 2.8e-15 [9]CH3O2+NO3→ HCHO+HO2+NO2 1.2e-12 [10]CH3COO+NO3→ CH3O2+NO2+CO2 4e-12 [10]oRO2+NO3→ NO2+HO2 1.2e-12 [10]obio+NO3→ NO2+HO2 1.2e-12 [10]oPAN+NO3→ NO2+HO2 4e-12 [10]oRN1+NO3→ 1.5*NO2 1.2e-12 [10]

C2H4+O3→ HCHO+0.12*HO2+0.13*H2+0.44*CO Ae−B/T ,A=9.1e-15,B=2580 [1]

C3H6+O3 → 0.53*HCHO +0.5*CH3CHO +0.31*CH3O2

+0.28*HO2 +0.15*OH +0.065*H2 +0.4*CO +0.7*CH4

Ae−B/T ,A=5.5e-15,B=1880 [1]

C5H8+O3→ 0.67*MAC +0.26*MVK +0.55*OH +0.07*C3H6

+0.8*HCHO +0.06*HO2 +0.05*CO +0.3*O3

Ae−B/T ,A=1.2e-14,B=2013 [6]

MAC+O3 → 0.8*MGLYOX +0.7*HCHO +0.215*OH+0.275*HO2 +0.2*CO +0.2*O3

Ae−B/T ,A=5.3e-15,B=2520 [6]

MVK+O3 → 0.82*MGLYOX +0.8*HCHO +0.04*CH3CHO+0.08*OH +0.06*HO2 +0.05*CO +0.2*O3

Ae−B/T ,A=4.3e-15,B=2016 [6]

APINEN+O3 → 1.27*CH3CHO +0.53*CH3COE +0.14*CO+0.62*oRO2 +0.42*HCHO +0.85*OH +0.1*HO2

Ae−B/T ,A=1.0e-15,B=736 [8]

BPINEN+O3 → 1.27*CH3CHO +0.53*CH3COE +0.14*CO+0.62*oRO2 +0.42*HCHO +0.85*OH +0.1*HO2

Ae−B/T ,A=1.0e-15,B=736 [8]

LIMONE+O3 → 1.27*CH3CHO +0.53*CH3COE +0.14*CO+0.62*oRO2 +0.42*HCHO +0.85*OH +0.1*HO2

Ae−B/T ,A=1.0e-15,B=736 [8]

TERPEN+O3 → 1.27*CH3CHO +0.53*CH3COE +0.14*CO+0.62*oRO2 +0.42*HCHO +0.85*OH +0.1*HO2

Ae−B/T ,A=1.0e-15,B=736 [8]

OCIMEN+O3 → 1.27*CH3CHO +0.53*CH3COE +0.14*CO+0.62*oRO2 +0.42*HCHO +0.85*OH +0.1*HO2

Ae−B/T ,A=1.0e-15,B=736 [8]

HUMULE+O3 → 1.27*CH3CHO +0.53*CH3COE +0.14*CO+0.62*oRO2 +0.42*HCHO +0.85*OH +0.1*HO2

Ae−B/T ,A=1.0e-15,B=736 [8]

CH3O2+NO→ HCHO+NO2+HO2 Ae−B/T ,A=4.2e-12,B=-180 [1]CH3COO+NO→ CH3O2+NO2+CO2 2e-11 [1]oRO2+NO→ NO2+HO2 4e-12 n3obio+NO→ 0.86*NO2+0.78*HO2+0.14*ISNI Ae−B/T ,A=1.4e-11,B=180 [6]oPAN+NO→ NO2+HO2 1.4e-11 n3oRN1+NO→ 1.5*NO2 4e-11 n3


Table A2. Reaction list of the reduced MELCHIOR2 chemical mechanism (continued)

Reactions Kinetic constants Ref.CH3O2+HO2→ CH3O2H Ae−B/T ,A=4.1e-13,B=-790 [1]

CH3COO+HO2→ 0.67*PPA+0.33*O3 Ae−B/T ,A=4.3e-13,B=-1040 [1]

oRO2+HO2→ oROOH Ae−B/T ,A=2.7e-13,B=-1000 n4obio+HO2→ obioH Ae−B/T ,A=2.7e-13,B=-1000 n4oPAN+HO2→ PANH Ae−B/T ,A=2.7e-13,B=-1000 n4oRN1+HO2→ X Ae−B/T ,A=2.7e-13,B=-1000 n4CH3O2+CH3O2→ 1.35*HCHO+0.7*HO2 Ae−B/T ,A=1.13e-13,B=-356 [1],n5CH3COO+CH3O2→ 0.5*CH3O2+0.5*CO2+HCHO+0.5*HO2 Ae−B/T ,A=3.34e-12,B=-400 [1],n5oRO2+CH3O2→ 0.65*HCHO+0.8*HO2+0.35*CH3OH Ae−B/T ,A=1.5e-13,B=-220 n5obio+CH3O2→ 0.8*HO2+0.5*HCHO Ae−B/T ,A=2.44e-11,B=223 n5oPAN+CH3O2→ HCHO+0.5*HO2 Ae−B/T ,A=7.9e-12,B=-140 n5CH3COO+CH3COO→ 2.*CH3O2+2.*CO2 Ae−B/T ,A=2.8e-12,B=-530 [1],n5oRO2+CH3COO→ 0.8*CH3O2+0.8*CO2+0.8*HO2 Ae−B/T ,A=8.6e-13,B=-260 n5obio+CH3COO→ 0.5*HCHO+1.5*HO2+0.7*CO2 Ae−B/T ,A=1.18e-11,B=127 n5oPAN+CH3COO→ CH3O2+CO2+HO2 Ae−B/T ,A=3.34e-12,B=-400 n5oRO2+oRO2→ 1.3*HO2 6.4e-14 n5CH3COO+NO2+M→ PAN troe(2.7e-28,0,7.1,1.2e-11,0,0.9,0.3) [1],n9oPAN+NO2+M→ toPAN troe(2.7e-28,0,7.1,1.2e-11,0,0.9,0.3) n6,n9PAN+M→ CH3COO+NO2 troe(4.9e-3,12100,0,5.4e16,13830,0,0.3) [1],n9toPAN+M→ oPAN+NO2 troe(4.9e-3,12100,0,5.4e16,13830,0,0.3) n6,n9PAN+OH→ HCHO+NO3+CO2 Ae−B/T ,A=9.5e-13,B=650 [1]

toPAN+OH→ NO3+CO2 Ae−B/T ,A=3.25e-13,B=-500 n6ISNI+OH → 0.95*CH3CHO +0.475*CH3COE +0.475*MG-LYOX +0.05*ISNI +0.05*HO2 + oRN1

3.4e-11 [6]

O3→ 2*OH photorate (for calculation, see §7.2) [3],n7NO2→ NO+O3 photorate [3]NO3→ NO2+O3 photorate [3], [11]NO3→ NO photorate [3], [11]N2O5→ NO2+NO3 photorate [3]H2O2→ 2*OH photorate [3]HNO3→ NO2+OH photorate [3]HONO→ NO+OH photorate [3]HCHO→ CO+2*HO2 photorate [3]HCHO→ CO+H2 photorate [3]CH3CHO→ CH3O2+HO2+CO photorate [1]CH3COE→ CH3COO+CH3CHO+oRO2 photorate [11], [12]CH3COY→ 2*CH3COO photorate [13], [14]MGLYOX→ CH3COO+HO2+CO photorate [1]GLYOX→ 0.6*HO2+2*CO+0.7*H2 photorate [13], [14]MEMALD→ 0.5*MVK +0.5*MALEIC +0.5*oPAN+0.5*HCHO +0.5*HO2

photorate [15]

CH3O2H→ HCHO+OH+HO2 photorate [3]PPA→CH3O2+CO2+OH photorate n8oROOH→OH+HO2 photorate n8obioH→ OH+HO2 photorate n8PANH→OH+HO2 photorate n8PAN→CH3COO+NO2 photorate [3]


The numbers in the Tables correspond to the following ref-erences:

1. Atkinson et al. (1997),

2. Donahue et al. (1997),

3. De Moore et al. (1994),2645

4. Mentel et al. (1996),

5. Atkinson (1990),

6. Paulson and Seinfeld (1992),

7. LeBras et al. (1997)

8. Atkinson (1994),2650

9. DeMore et al. (1990)

10. Canosa-Mas et al. (1996),

11. Johnston et al. (1996),

12. Martinez et al. (1992),

13. Jenkin et al. (1997),2655

14. Plum et al. (1983),

15. Bierbach et al. (1994),

16. Lightfoot and Cox (1992),

17. Kirchner and Stockwell (1996),

18. Aumont et al. (2003)2660

The notes for the reactions correspond to the following re-marks:

• n1: Rate constants containing a density dependent term ina minor reaction channel have been simplified by puttingan average boundary layer molecular density of 2.5 1019

2665

molecules/cm3.• n2: For operator reactions with OH, oROOH is inter-

preted as 2-butyl hydrogen peroxide, obioH as a peroxidefrom isoprene degradation, PANH represents unsaturatedorganic peroxides; due to lack of data, the rate constants2670

of these compounds and of peroxy acetyl acid have beenestimated using structure reactivity relationships in Kwokand Atkinson (1995).

• n3: The rate constant of the operator reaction oRO2+NOhas been set to that of the i-C3H7O2+NO reaction given2675

in Lightfoot and Cox (1992), that of oPAN+NO to thatof CH3COO2+NO [1], the rate constant for oRO2NO+NOis taken from that of ISNIR (nitrate peroxy radicals fromisoprene degradation) + NO, Paulson and Seinfeld (1992).

• n4: The rate constants for the operator reactions oRO2,2680

obio, RO2NI+HO2 have been set to that of the reac-tion C2H5O2+HO2 and oPAN+HO2 has been taken fromCH3COO2+HO2 as in Atkinson et al. (1997).• n5: Radical conserving and terminating pathways are

combined to single reactions; rate constants in operator2685

reactions are affected by interpreting oRO2 as 2-butyl per-oxy radical Lightfoot and Cox (1992), obio as a peroxyradical from isoprene degradation Paulson and Seinfeld(1992) and oPAN as CH3COO2.• n6: Reaction rates of the operators oPAN and toPAN have2690

been set to those of the peroxy acetyl radical and PAN,respectively.• n7: The reactions O3+hν→ O(1D), O(1D)+H2O→ 2OH,

O(1D)+M→ O(3P)+M and O(3P)+O2+M→ O3+M havebeen combined to reaction O3+hν→ O(1D) by taking into2695

account H2O and M concentrations in order to adjust thephotolysis frequency.• n8: Due to lack of data, photolysis frequencies of higher

organic peroxides are taken as that of CH3OOH.• n9: Three body Troe reactions, given in the form, as in2700

Atkinson et al. (1997):

k=k0[M ]

1+k0[M ]k∞

fp (A1)

with

p= (1+(log10(k0[M ]/k∞))2)−1, (A2)

k0 =A0e

0@−B0

T

1A(T

300

)−n, (A3)2705

k∞=A∞e

0@−B∞T

1A(T

300

)−n, (A4)

The parameters are given in the orderA0,B0,n,A∞,B∞,m,f

References

Akagi, S. K., Yokelson, R. J., Wiedinmyer, C., Alvarado, M. J.,2710

Reid, J. S., Karl, T., Crounse, J. D., and Wennberg, P.: Emissionfactors for open and domestic biomass burning for use in atmo-spheric models, Atmos. Chem. Phys., 11, 4039–4072, 2011.

Alfaro, S. C. and Gomes, L.: Modeling mineral aerosol productionby wind erosion: Emission intensities and aerosol size distribu-2715

tion in source areas, J. Geophys. Res., 106, 18,075–18,084, 2001.Allen, A., Nemitz, E., Shi, J., Harrison, R., and Greenwood, J.: Size

distribution of trace metals in atmospheric aerosols in the UnitedKingdom, Atmospheric Environment, 35, 4581–4591, 2001.

Alvarado, M. J. and Prinn, R. G.: Formation of ozone and growth2720

of aerosols in young smoke plumes from biomass burning, Part1: Lagrangian parcel studies, J. Geophys. Res., 114, D09 306,doi:10.1029/2008JD011144, 2009.


Anav, A., Menut, L., Khvorostiyanov, D., and Viovy, N.: Im-pact of tropospheric ozone on the Euro-Mediterranean vegeta-2725

tion, Global Change Biology, 17, 2011.Atkinson, R.: Gas-phase tropospheric chemistry of organic-

compounds - a review, Atmospheric Environment, 24, 1 – 41,1990.

Atkinson, R.: Gas-phase tropospheric chemistry of organic-2730

compounds, Journal of Physical and Chemical reference data,p. R1, 1994.

Atkinson, R., Baulsch, D. L., Cox, R. A., Hampton, R. F., Kerr,J. A., Rossi, M. J., and Troe, J.: Evaluated kinetics, photochemi-cal and heterogeneous data, J. Phys. Chem., 26, 521–1012, 1997.2735

Attrep, M. and Anirudhan, M.: Atmospheric inorganic and organicarsenic, Trace Subst. Environ. Health, 11, 365–369, 1977.

Aumont, B., Chervier, F., and Laval, S.: Contribution of HONOto the NOx/HOx/O3 chemistry in the polluted boundary layer,Atmos. Environ., 37, 487–498, 2003.2740

Aumont, B., Szopa, S., and Madronich, S.: Modelling the evolutionof organic carbon during its gas-phase tropospheric oxidation:development of an explicit model based on a self generating ap-proach, Atmos. Chem. Phys., pp. 2497–2517, 2005.

Baer, M. and Nester, K.: Parameterization of trace gas dry depo-2745

sition velocities for a regional mesoscale diffusion model., Ann.Geophys., 10, 912–923, 1992.

Berge, E.: Coupling of wet scavenging of sulphur to clouds in anumerical weather prediction model, Tellus, 45B, 1–22, 1993.

Bessagnet, B., Hodzic, A., Vautard, R., Beekmann, M., Cheinet,2750

S., Honore, C., Liousse, C., and Rouil, L.: Aerosol modelingwith CHIMERE: preliminary evaluation at the continental scale,Atmospheric Environment, 38, 2803–2817, 2004.

Bessagnet, B., Hodzic, A., Blanchard, O., Lattuati, M., Le Bi-han, O., and Marfaing, H.: Origin of particulate matter pollution2755

episodes in wintertime over the Paris Basin, Atmospheric Envi-ronment, 39, 6159–6174, 2005.

Bessagnet, B., Menut, L., Aymoz, G., Chepfer, H., and Vautard,R.: Modelling dust emissions and transport within Europe: theUkraine March 2007 event, Journal of Geophysical Research,2760

113, D15 202, doi:10.1029/2007JD009541, 2008.Bessagnet, B., Menut, L., Curci, G., Hodzic, A., Guillaume, B., Li-

ousse, C., Moukhtar, S., Pun, B., Seigneur, C., and Schulz, M.:Regional modeling of carbonaceous aerosols over Europe - Fo-cus on Secondary Organic Aerosols, Journal of of Atmospheric2765

Chemistry, 61, 175–202, 2009.Bicheron, P., Amberg, V., Bourg, L., Petit, D., Huc, M., Miras,

B., Brockmann, C., Hagolle, O., Delwart, S., Ranera, F., Leroy,M., and Arino, O.: Geolocation Assessment of MERIS Glob-Cover Orthorectified Products, IEEE TRANSACTIONS ON2770

GEOSCIENCE AND REMOTE SENSING, 49, 2972–2982, doi:{10.1109/TGRS.2011.2122337}, 2011.

Bierbach, A., Barnes, I., Becker, K. H., and Wiesen, E.: Atmo-spheric chemistry of unsaturated carbonyls - butenedial, 4-oxo-2-pentenal, 3-hexene-2,5-dione, maleic-anhydride, 3H-furan-2-2775

one, and 5-mehtyl-3H-furan-2-one, Environ. Sci. Technol., 28,715 – 729, 1994.

Blond, N. and Vautard, R.: Three-dimensional ozone analyses andtheir use for short-term ozone forecasts, JOURNAL OF GEO-PHYSICAL RESEARCH-ATMOSPHERES, 109, doi:{10.1029/2780

2004JD004515}, 2004.Blond, N., Bel, L., and Vautard, R.: Three-dimensional ozone data

analysis with an air quality model over the Paris area, JOURNALOF GEOPHYSICAL RESEARCH-ATMOSPHERES, 108, doi:{10.1029/2003JD003679}, 2003.2785

Blond, N., Boersma, K. F., Eskes, H. J., van der A, R. J.,Van Roozendael, M., De Smedt, I., Bergametti, G., and Vautard,R.: Intercomparison of SCIAMACHY nitrogen dioxide observa-tions, in situ measurements and air quality modeling results overWestern Europe, JOURNAL OF GEOPHYSICAL RESEARCH-2790

ATMOSPHERES, 112, doi:{10.1029/2006JD007277}, 2007.Boichu, M., Menut, L., Khvorostyanov, D., Clarisse, L., Clerbaux,

C., Turquety, S., and Coheur, P.-F.: Inverting for volcanic SO2flux at high temporal resolution using spaceborne plume im-agery and chemistry-transport modelling: the 2010 Eyjafjalla-2795

jokull eruption case-study, Atmos. Chem. Phys. Discuss., 13,6553–6588, 2013.

Bowman, F. m., Odum, J. R., Seinfeld, J. H., and Pandis, S. N.:Mathematical model for gas-particle parttioning of secondary or-ganic aerosols, Atmos. Environ., 31, 3921–3931, 1997.2800

Broquet, G., Chevallier, F., Rayner, P., Aulagnier, C., Pison, I., Ra-monet, M., Schmidt, M., Vermeulen, A. T., , and Ciais, P.: AEuropean summertime CO2 biogenic flux inversion at mesoscalefrom continuous in situ mixing ratio measurements, J. Geophys.Res., 116, D23 303, doi:10.1029/2011JD016202, 2011.2805

Burney, P., Chinn, S., Jarvis, D., Luczynska, C., and Lai, E.: Vari-ations in the prevalence of respiratory symptoms, self-reportedasthma attacks, and use of asthma medication in the EuropeanCommunity Respiratory Health Survey (ECRHS), Eur. Respir.J., 9, 687–695, 1996.2810

Byun, D. W., Young, J., Pleim, J., Odman, M. T., and Alapaty, K.:Numerical transport algorithms for the community multiscale airquality (CMAQ) chemical transport model in generalized coor-dinates, in: Science Algorithms of the EPA Models-3 Commu-nity Multiscale Air Quality (CMAQ) Modeling System, US-EPA2815

Office of Research and Development Washington, DC 20460,EPA/600/R-99/030, 1999.

Cakmur, R. V., Miller, R. L., and Torres, O.: Incorporating the effectof small-scale circulations upon dust emission in an atmosphericgeneral circulation model, J. Geophys. Res., 109, D07 201, doi:2820

10.1029/2003JD004067, 2004.Canosa-Mas, C. E., King, M. D., Lopez, R., Percival, C. J., Wayne,

R. P., Pyle, J. A., Shallcross, D. E., and Daele, V.: Is the reac-tion between CH3C(O)O2 and NO3 important in the night-timetroposphere?, J. Chem. Soc. Faraday Trans., 92, 4385 – 4389,2825

1996.Carter, W.: Development of the SAPRC-07 chemical mechanism,

Atmospheric Environment, 44, 5324 – 5335, doi:10.1016/j.atmosenv.2010.01.026, 2010.

Carter, W. P. L.: A detail mechanism for the gas-phase atmospheric2830

reactions of organic compounds, Atmospheric Environment, 24,481–518, 1990.

Carter, W. P. L.: Documentation of the SAPRC-99 chemical mech-anism for VOC reactivity assessment: Final report to CaliforniaAir Resources Board, Contract 92-329 and Contract 95-308, Cal-2835

ifornia Air Resources Board, Sacramento, Calif., 2000.Chaxel, E., Rieux, C., Rios, I., Thibaudon, M., and Oliver, G.: Mod-

eling the 2011 ragweed season in the French region Rhone-Alpeswith the dispersion model CHIMERE, in: International RagweedConference, Lyon, France, March, 28 2012, 2012.2840

Cheinet, S. and Teixeira, J.: A simple formulation for the eddy-


diffusivity parameterization of cloud-topped boundary layers,GEOPHYSICAL RESEARCH LETTERS, 30, doi:{10.1029/2003GL017377}, 2003.

Christensen, J., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held,2845

I., Jones, R., Kolli, R., Kwon, W.-T., Laprise, R., Rueda, V. M.,Mearns, L., Menendez, C., Raisanen, J., Rinke, A., Sarr, A., andWhetton, P.: Regional Climate Projections. Chapter 11 of: Cli-mate Change 2007: The Physical Science Basis. Contribution ofWorking Group I to the Fourth Assessment Report of the Inter-2850

governmental Panel on Climate Change (S. Solomon et al., eds),Cambridge University Press, Cambridge, United Kingdom andNew York, NY, USA, 2007.

Clerbaux, C., Boynard, A., Clarisse, L., George, M., Hadji-Lazaro,J., Herbin, H., Hurtmans, D., Pommier, M., Razavi, A., Turquety,2855

S., Wespes, C., and Coheur, P.-F.: Monitoring of atmosphericcomposition using the thermal infrared IASI/MetOp sounder, At-mos. Chem. Phys., 9, 6041–6054, 2009.

Colella, P. and Woodward, P. R.: The piecewise parabolic method(PPM) for gas-dynamical simulations, Journal of Computational2860

Physics, 11, 38–39, 1984.Colette, A., Favez, O., Meleux, F., Chiappini, L., Haeffelin, M.,

Morille, Y., Malherbe, L., Papin, A., Bessagnet, B., Menut, L.,Leoz, E., and Rouıl, L.: Assessing in near real time the impact ofthe April 2010 Eyjafjallajokull ash plume on air quality, Atmo-2865

spheric Environment, 45, 1217–1221, 2011a.Colette, A., Granier, C., Hodnebrog, Ø., Jakobs, H., Maurizi, A.,

Nyiri, A., Bessagnet, B., D’Angiola, A., D’Isidoro, M., Gauss,M., Meleux, F., Memmesheimer, M., Mieville, A., Rouıl, L.,Russo, F., Solberg, S., Stordal, F., and Tampieri, F.: Air quality2870

trends in Europe over the past decade: a first multi-model assess-ment, Atmospheric Chemistry and Physics, 11, 11 657–11 678,doi:10.5194/acp-11-11657-2011, 2011b.

Colette, A., Meleux, F., Bessagnet, B., Granier, C., Hodnebrog,A., Pirovano, G., and Szopa, S.: On the impact of chem-2875

ical boundary conditions on air quality modelling, in: Geo-physical Research Abstracts, edited by EGU, vol. 13 of , pp.EGU2011–11 655, EGU annual meeting at Vienna, Austria, doi:{10.1007/978-1-4020-6766-2\ 6}, 2011.

Coman, C., Foret, G., Beekmann, M., Eremenko, M., Dufour, G.,2880

Gaubert, B., Ung, A., Schmechtig, C., Flaud, J.-M., , and Berga-metti, G.: Assimilation of IASI partial tropospheric columnswith an Ensemble Kalman Filter over Europe, Atmos. Chem.Phys., 12, 2513–2532, 2012.

COSY: Experimental regional forecast at Laboratoire de Mete-2885

olologie Dynamique, with WRF and CHIMERE. Daily compar-isons to the SIRTA observatory measurements., http://www.lmd.polytechnique.fr/cosy/,lastaccessMarch2013, 2013.

Courant, R., Isaacson, E., and Rees, M.: On the solution ofnonlinear hyperbolic differential equations by finite differences,2890

Comm. Pure and Applied Mathematics, 5, 243–255, 1952.Curci, G., Beekmann, M., Vautard, R., Smiatek, G., Steinbrecher,

R., Theloke, J., and Friedrich, R.: Modelling study of the impactof isoprene and terpene biogenic emissions on European ozonelevels, Atmospheric Environment, 43, 1444–1455, doi:10.1016/2895

j.atmosenv.2008.02.070, 2009.Curci, G., Palmer, P. I., Kurosu, T. P., Chance, K., and Visconti, G.:

Estimating European volatile organic compound emissions usingsatellite observations of formaldehyde from the Ozone Monitor-ing Instrument, Atmos. Chem. Phys., 10, 11 501–11 517, 2010.2900

Cuvelier, C., Thunis, P., Vautard, R., Amann, M., Bessagnet, B.,Bedogni, M., Berkowicz, R., Brocheton, F., Builtjes, P., Denby,B., Douros, G., Graf, A., Honore, C., Jonson, J., Kerschbaumer,A., de Leeuw, F., Moussiopoulos, N., Philippe, C., Pirovano, G.,Rouil, L., Schaap, M., Stern, R., Tarrason, L., Vignati, E., Volta,2905

L., White, L., Wind, P., and Zuber, A.: CityDelta: a model inter-comparison study to explore the impact of emission reductionsin European cities in 2010, Atmos. Environ., 41, 189–207, doi:10.1016/j.atmosenv.2006.07.036, 2007.

Darbah, J., Kubiske, M., Nelson, N., Oksanen, E., and Vapaavuori,2910

E.: Effects of decadal exposure to interacting elevated CO2and/or O3 on paper birch (Betula papyrifera) reproduction, Env-iron Pollut, 155, 446–452, 2008.

de Meij, A., Gzella, A., Cuvelier, C., Thunis, P., Bessagnet, B.,Vinuesa, J., Menut, L., and H., K.: The impact of MM5 and2915

WRF meteorology over complex terrain on CHIMERE modelcalculations, Atmos. Chem. Phys., 9, 6611–6632, 2009.

De Moore, W. B., Sandetr, S. P., Golden, D. M., Hampton, R. F.,Kurylo, M. J., Howard, C. J., Ravishankara, A. R., Kolb, C. E.,and Molina, M. J.: Chemical kinetics and photochimical data for2920

use in stratospheric modelling evaluation, JPL publication, 94,26, JPL, Pasadena, US, 1994.

Deguillaume, L., Beekmann, M., and Derognat, C.: Uncertaintyevaluation of ozone production and its sensitivity to emis-sion changes over the Ile-deFrance region during summer pe-2925

riods, Journal of Geophysical Research, 113, doi:doi:10.1029/2007JD009081, 2008.

DeMore, W., Sander, S., Golden, D., Molina, M., Hampson, R.,Kurylo, M., Howard, C., and Ravishankara, A.: Chemical ki-netics and photochemical data for use in stratospheric modeling.2930

Evaluation number 9., Jet Propulsion Lab., California Inst. Tech-nol., Pasadena, CA, 1990.

Derognat, C., Beekmann, M., Baeumle, M., Martin, D., andSchmidt, H.: Effect of biogenic volatile organic compound emis-sions on tropospheric chemistry during the Atmospheric Pol-2935

lution Over the Paris Area (ESQUIF) campaign in the Ile-de-France region, JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES, 108, doi:{10.1029/2001JD001421}, 2003.

Donahue, N. M., Dubey, M. K., Mohrschladt, R., Demerjian, K. L.,and Anderson, J. G.: High-pressure flow study of the reactions2940

OH+NOx-¿HONOx: Errors in the falloff region, Journal of Geo-physical Research, 102, 6159 – 6168, 1997.

Dufour, G., Wittrock, F., Camredon, M., Beekmann, M., Richter,A., and Aumont, B.: SCIAMACHY formaldehyde observations:constraint for isoprene emission estimates over Europe?, Atmos.2945

Chem. Phys., 9, 1647–1664, 2009.Dufour, G., Eremenko, M., Griesfeller, A., Barret, B.,

LeFlochmoen, E., Clerbaux, C., Hadji-Lazaro, J., Coheur, P.-F.,and Hurtmans, D.: Validation of three different scientific ozoneproducts retrieved from IASI spectra using ozonesondes, Atmos.2950

Meas. Tech., 5, 611–630, 2012.ECCAD: Emissions of atmospheric Com-

pounds and Compilation of Ancillary Data,http://eccad.sedoo.fr/eccad extract interface/JSF/page login.jsf.

Elbern, H. and Schmidt, H.: Ozone episode analysis by four dimen-2955

sional variational chemistry data assimilation, J. Geophys. Res.,106, 3569– 3590, 2001.

EPA: National Air Toxics Program: The Integrated Urban Strat-egy, United States Environmental Protection Agency’s Federal


Register, US EPA Reports, p. Last updated on 07/19/2012, doi:2960

www.epa.gov/ttn/uatw/urban/urbanpg.html, 1999.Eremenko, M., Dufour, G., Foret, G., Keim, C., Orphal, J., Beek-

mann, M., Bergametti, G., and Flaud, J. M.: Troposphericozone distributions over Europe during the heat wave in July2007 observed from infrared nadir spectra recorded by IASI,2965

GEOPHYSICAL RESEARCH LETTERS, 35, doi:{10.1029/2008GL034803}, 2008.

Erisman, J. W., van Pul, A., and Wyers, P.: Parameterization of sur-face resistance for the quantification of atmospheric depositionof acidifying pollutants and ozone, Atmospheric Environemnt,2970

28, 2595–2607, 1994.Evensen, G.: Sequential data assimilation with a nonlinear quasi-

geostrophic model using Monte Carlo methods to forecast errorstatistics, J. Geophys. Res., 99, 143–162, 1994.

Fenger, J.: Air pollution in the last 50 years - From local to global,2975

Atmospheric Environment, 43, 13–22, 2009.Finlayson-Pitts, B. J. and Pitts, J.: Theory, Experiments, and Appli-

cations, in: Chemistry of the Upper and Lower Atmosphere (FirstEdition), San Diego: Academic Press, ISBN:978-0-12-257060-5, 2000.2980

Folberth, G. A., Hauglustaine, D. A., Lathiere, J., and Brocheton,F.: Interactive chemistry in the Laboratoire de Meteorologie Dy-namique general circulation model: model description and im-pact analysis of biogenic hydrocarbons on tropospheric chem-istry, Atmos. Chem. Phys., 6, 2273–2319, 2006.2985

Freitas, S. R., Longo, K. M., Chatfield, R., Latham, D., Silva Dias,M. A. F., Andreae, M. O., Prins, E., Santos, J. C., Gielow, R., ,and Carvalho Jr., J. A.: Including the sub-grid scale plume rise ofvegetation fires in low resolution atmospheric transport models,Atmos. Chem. Phys., 7, 3385– 3398, 2007.2990

Freitas, S. R., Rodrigues, L. F., Longo, K. M., and Panetta, J.: Im-pact of a monotonic advection scheme with low numerical diffu-sion on transport modeling of emissions from biomass burning,JOURNAL OF ADVANCES IN MODELING EARTH SYS-TEMS, 3, doi:{10.1029/2011MS000084}, 2011.2995

Friedrich, R.: GENEMIS: Generation of European Emission Datafor Episodes, in: Transport and Chemical Transformation of Pol-lutants in the Troposphere, edited by Borrell, P. and Borrell,P., vol. 1 of Transport and Chemical Transformation of Pollu-tants in the Troposphere, pp. 375–386, Springer Berlin Heidel-3000

berg, doi:10.1007/978-3-642-59718-3 18, http://dx.doi.org/10.1007/978-3-642-59718-3 18, 2000.

Fuchs, N.: The Mechanics of Aerosols, Pergamon Press, Oxford,1964.

Gelbard, F. and Seinfeld, J. H.: Simulation of multicomponent3005

aerosol dynamics, Journal of colloid and Interface Science, 78,485–501, 1980.

Giglio, L., Randerson, J. T., van der Werf, G. R., Kasibhatla, P. S.,Collatz, G. J., Morton, D. C., and DeFries, R. S.: Assessing vari-ability and long-term trends in burned area by merging multi-3010

ple satellite fire products, Biogeosciences, 7, 1171–1186, doi:10.5194/bg-7-1171-2010, 2010.

Ginoux, P., Chin, M., Tegen, I., Prospero, J. M., Holben, B.,Dubovik, O., and Lin, S. J.: Sources and distributions of dustaerosols simulated with the GOCART model, Journal of Geo-3015

physical Research, 106, 20 255–20 273, 2001.Giorgi, F.: A particle dry deposition scheme for use in tracer trans-

port models‘, J. Geophys. Res., 91, 9794–9806, 1986.

Gonzalez, M., Vivanco, M. G., Palomino, I., Garrido, J., Santiagao,M., and Bessagnet, B.: Modelling some heavy metals air con-3020

centration in Europe, Water Air and Soil Pollution, p. in press,2012.

Griffin, R. J., Cocker, E. R., Flagan, R. C., and Seinfeld, J. H.: Or-ganic aerosol formation from the oxidation of biogenic hydrocar-bons, Journal of Geophysical Research, 104, 3555–3567, 1999.3025

Guelle, W., Balkanski, Y. J., Dibb, J. E., Schulz, M., and Dulac,F.: Wet deposition in a global size-dependent aerosol transportmodel. 2. Influence of the scavenging scheme on Pb verticalprofiles, and deposition, J. Geophys. Res., 103, 28 875–28 891,1998.3030

Guenther, A., Karl, T., Harley, P., Wiedinmyer, C., Palmer, P., andGeron, C.: Estimates of global terrestrial isoprene emissions us-ing MEGAN (Model of Emissions of Gases and Aerosols fromNature), Atmos. Chem. Phys., 6, 3181–3210, 2006.

Hall, S., Khudaish, E., and Hart, A.: Electrochemical oxidation of3035

hydrogen peroxide at platinum electrodes. Part III: Effect of tem-perature. Electrochim, Electrochim. Acta, 44, 2455–2462, 1999.

Hamaoui-Laguel, L., Meleux, F., Beekmann, M., Bessagnet, B.,Genermont, S., Cellier, P., and Letinois, L.: Improving ammo-nia emissions in air quality modelling for France, Atmospheric3040

Environment, pp. –, doi:10.1016/j.atmosenv.2012.08.002, 2012.Hanea, R. G., Velders, G. J. M., and Heemink, A.: Data assimilation

of ground-level ozone in Europe with a Kalman filter and chem-istry transport model, Journal of Geophysical Research: Atmo-spheres, 109, doi:10.1029/2003JD004283, http://dx.doi.org/10.3045

1029/2003JD004283, 2004.Hansen, M. C. and Reed, B.: A comparison of the IGBP DISCover

and University of Maryland 1 km global land cover products,International Journal of Remote Sensing, 21, 1365–1373, 2000.

Harrison, R. and Kito, A.: Field intercomparison of filter pack and3050

denuder sampling methods for reactive gaseous and particulatepollutants, Atmos. Environ., 24, 2633–2640, 1990.

HEI: Summary of a Workshop on Metal-Based Fuel Additives andNew Engine Technologies, 1998.

Hodzic, A.: Modelisation des aerosols de pollution en Ile de France,3055

Ph.D. thesis, Ecole Polytechnique, 2005.Hodzic, A. and Jimenez, J. L.: Modeling anthropogenically con-

trolled secondary organic aerosols in a megacity: a simplifiedframework for global and climate models, Geoscientific ModelDevelopment, 4, 901–917, doi:10.5194/gmd-4-901-2011, 2011.3060

Hodzic, A., Vautard, R., Bessagnet, B., Lattuati, M., and Moreto,F.: Long-term urban aerosol simulation versus routine particulatematter observations, Atmospheric Environemnt, 39, 5851–5864,2005.

Hodzic, A., Bessagnet, B., and Vautard, R.: A model evaluation of3065

coarse-mode nitrate heterogeneous formation on dust particles,Atmos. Environ., 40, 4158–4171, 2006a.

Hodzic, A., Vautard, R., Chazette, P., Menut, L., and B.Bessagnet:Aerosol chemical and optical proprties over the Paris area withinESQUIF project, Atmospheric Chemistry and Physics, 6, 3257–3070

3280, 2006b.Hodzic, A., Vautard, R., Chepfer, H., Goloub, P., Menut, L.,

Chazette, P., Deuze, J., Apituley, A., and Couvert, P.: Evolu-tion of aerosol optical thickness over Europe during the August2003 heat wave as seen from CHIMERE model simulations and3075

POLDER data, Atmospheric Chemistry and Physics, 6, 1853–1864, 2006c.


Hodzic, A., Madronich, S., Bohn, B., Massie, S., Menut, L., , andWiedinmyer, C.: Wildfire particulate matter in Europe duringsummer 2003: Meso-scale modeling of smoke emissions, trans-3080

port and radiative effects, Atmospheric Chemistry and Physics,7, 2007.

Hodzic, A., Jimenez, J. L., Madronich, S., Aiken, A. C., Bessagnet,B., Curci, G., Fast, J., Lamarque, J.-F., Onasch, T. B., Roux, G.,Schauer, J. J., Stone, E. A., and Ulbrich, I. M.: Modeling organic3085

aerosols during MILAGRO: importance of biogenic secondaryorganic aerosols, Atmospheric Chemistry and Physics, 9, 6949–6981, doi:10.5194/acp-9-6949-2009, 2009.

Hodzic, A., Jimenez, J., Madronich, S., Canagaratna, M., DeCarlo,P., Kleinman, L., and Fast, J.: Modeling organic aerosols in3090

a megacity: potential contribution of semi-volatile and inter-mediate volatility primary organic compounds to secondary or-ganic aerosol formation, Atmos. Chem. Phys., 10, doi:10.5194/acp-10-5491-2010, 2010a.

Hodzic, A., Jimenez, J., Prevot, A., Szidat, S., Fast, J., and3095

Madronich, S.: Can 3D Models Explain the Observed Fractionsof Fossil and non-Fossil Carbon In and Near Mexico City?, At-mos. Chem. Phys., 10, doi:10.5194/acp-10-10997-2010, 2010b.

Hodzic, A., Chepfer, H., Vautard, R., Chazette, P., Beekmann,M., Bessagnet, B., Chatenet, B., Cuesta, J., Drobinski, P.,3100

Goloub, P., Haeffelin, M., and Morille, Y.: Comparison ofaerosol chemistry transport model simulations with lidar andSun photometer observations at a site near Paris, JOURNALOF GEOPHYSICAL RESEARCH-ATMOSPHERES, 109, doi:{10.1029/2004JD004735}, 2004.3105

Hoffmann, M. R. and Calvert, J. G.: Chemical transformationmodules for eulerian acid deposition models,, ePA/600/3-85/017,1985.

Hollingsworth, A., Engelen, R. J., Benedetti, A., Dethof, A., Flem-ming, J., Kaiser, J. W., Morcrette, J.-J., Simmons, A. J., Textor,3110

C., Boucher, O., Chevallier, F., Rayner, P., Elbern, H., Eskes, H.,Granier, C., Peuch, V.-H., Rouil, L., and Schultz, M. G.: Towarda monitoring and forecasting system for atmospheric composi-tion: The GEMS project, B. Am. Meteorol. Soc., 89, 1147–1164,2008.3115

Honore, C. and Vautard, R.: Photochemical regimes in urban atmo-spheres: The influence of dispersion, Geophys. Res. Letters, 27,1895–1898, 2000.

Honore, C., Rouıl, L., Vautard, R., Beekmann, M., Bessagnet, B.,Dufour, A., Elichegaray, C., Flaud, J., Malherbe, L., Meleux, F.,3120

Menut, L., Martin, D., Peuch, A., Peuch, V., and Poisson, N.:Predictability of European air quality: The assessment of threeyears of operational forecasts and analyses by the PREV’AIRsystem, Journal of Geophysical Research, 113, D04 301, doi:10.1029/2007JD008761, 2008.3125

Hourdin, F. and Armengaud, A.: On the use of finite volume meth-ods for atmospheric advection of trace species. Part I: Test ofvarious formulations in a General Circulation Models, MonthlyWeather Review, 127, 822–837, 1999.

Huijnen, V., Eskes, H. J., Poupkou, A., Elbern, H., Boersma, K. F.,3130

Foret, G., Sofiev, M., Valdebenito, A., Flemming, J., Stein, O.,Gross, A., Robertson, L., D’Isidoro, M., Kioutsioukis, I., Friese,E., Amstrup, B., Bergstrom, R., Strunk, A., Vira, J., Zyryanov,D., Maurizi, A., Melas, D., Peuch, V.-H., and Zerefos, C.: Com-parison of OMI NO2 tropospheric columns with an ensemble of3135

global and European regional air quality models, Atmos. Chem.

Phys., 10, 3273–3296, 2010.Huneeus, N., Chevallier, F., , and Boucher, O.: Estimating aerosol

emissions by assimilating observed aerosol optical depth in aglobal aerosol model, Atmos. Chem. Phys., 12, 4585–4606,3140

2012.Jacob, D. J.: Heterogeneous chemistry and tropospheric ozone, At-

mospheric Environment, 34, 2131–2159, 2000.Jaffe, D. A. and Wigder, N. L.: Ozone production from wildfires

: a critical review, Atmospheric Environment, 51, 1–10, doi:10.3145

1016/j.atmosenv.2011.11.063, 2012.Jenkin, M. E., Saunders, S. M., and Pilling, M.: The tropospheric

degradation of volatile organic compounds: A protocol for mech-anism development., Atmospheric Environment, 31, 81–104,1997.3150

Johnson, D. L. and Braman, R. S.: Alkyl- and inorganic arsenic inair samples, Chemosphere, 6, 333–338, 1975.

Johnston, H. S., Davis, H. F., and Lee, Y. T.: NO3 photolysis prod-uct channels: Quantum yields from observed energy thresholds,J. Phys. Chem., 100, 4713 – 4723, 1996.3155

Jung, C., Kim, Y., and Lee, K.: Analytic solution for polydispersedaerosol dynamics by a wet removal process., Journal of AerosolScience, 33, 753–767, 2002.

Kaiser, J. W., Heil, A., Andreae, M. O., Benedetti, A., Chubarova,N., Jones, L., Morcrette, J.-J., Razinger, M., Schultz, M. G.,3160

Suttie, M., and van der Werf, G. R.: Biomass burning emis-sions estimated with a global fire assimilation system based onobserved fire radiative power, Biogeosciences, 9, 527–554, doi:10.5194/bg-9-527-2012, 2012.

Kavouras, I., Mihalopoulos, N., and Stephanou, E.: Formation of3165

atmospheric particles from organic acids produced by forests,Nature, 395, 683–686, 1998.

Khvorostyanov, D. V., Menut, L., Dupont, J.-C., Morille, Y., andHaeffelin, M.: The role of WRF land surface schemes on weathersimulations in Paris area, in: Proceedings of ISARS 2010 con-3170

ference, Guyancourt, France, june, 28 2010, abstract number:O-SUR/11, 2010.

Kirchner, F. and Stockwell, W. R.: Effect of peroxy radical re-actions on the predicted concentrations of ozone, nitrogenouscompounds, and radicals, Journal of Geophysical Research, 101,3175

21 007 – 21 022, 1996.Konovalov, I. B., Beekmann, M., Vautard, R., Burrows, J., Richter,

A., Nuss, H., and Elansky, N.: Comparison and evaluation ofmodelled and GOME measurement derived tropospheric NO2

columns over Western and Eastern Europe, Atmospheric Chem-3180

istry and Physics, 5, 169–190, 2005.Konovalov, I. B., Beekmann, M., Richter, A., and Burrows, J.: In-

verse modelling of the spatial distribution of NOx emissions ona continental scale using satellite data, Atmospheric Chemistryand Physics, 6, 1747–1770, 2006.3185

Konovalov, I. B., Beekmann, M., Kuznetsova, I. N., Yurova, A.,and Zvyagintsev, A. M.: Atmospheric impacts of the 2010 Rus-sian wildfires: integrating modelling and measurements of an ex-treme air pollution episode in the Moscow region, AtmosphericChemistry and Physics, 11, 10 031–10 056, 2011.3190

Konovalov, I. B., Beekmann, M., D’Anna, B., and George, C.: Sig-nificant light induced ozone loss on biomass burning aerosol:Evidence from chemistry-transport modeling based on new lab-oratory studies, GEOPHYSICAL RESEARCH LETTERS, 39,doi:{10.1029/2012GL052432}, 2012.3195


Krinner, G., Viovy, N., de Noblet-Ducoudre, N., Ogee, J., Polcher,J., Friedlingstein, P., Ciais, P., Sitch, S., and Prentice, I.: Adynamic global vegetation model for studies of the coupledatmosphere-biosphere system, GLOBAL BIOGEOCHEMICALCYCLES, 19, doi:{10.1029/2003GB002199}, 2005.3200

Kroll, J. H., Ng, N. L., Murphy, S. M., Flagan, R. C., and Se-infeld, J. H.: Secondary organic aerosol formation from iso-prene photooxidation, Environ. Sci. Technol., 40, 1869–1877,doi:10.1021/es0524301, 2006.

Kuenen, J., Denier van der Gon, H., Visschedijk, A., and van der3205

Brugh, H.: High resolution European emission inventory forthe years 2003-2007, in: TNO report TNO-060-UT-2011-00588,Utrecht, 2011, 2011.

Kulmala, M., A., L., and Pirjola, L.: Parameterization for sulfuricacid / water nucleation rates, J. Geophys. Res., 103, 8301–8307,3210

1998.Kuloglu, E. and Tuncel, G.: Size distribution of trace elements and

major ions in the eastern Mediterranean atmosphere, Water, Air,and Soil Pollution, 167, 221–241, 2005.

Kurtenbach, R., Becker, K., Gomes, J., Kleffmann, J., Lorzer, J.,3215

Spittler, M., Wiesen, P., Ackermann, R., Geyer, A., and Platt,U.: Investigations of emissions and heterogeneous formation of{HONO} in a road traffic tunnel, Atmospheric Environment, 35,3385 – 3394, doi:10.1016/S1352-2310(01)00138-8, 2001.

Kwok, E. and Atkinson, R.: Estimation of hydroxyl radical3220

reaction-rate constants for gas-phase organic-compounds usinga structure-reactivity relationship - an update., Atmospheric En-vironment, 29, 1685–1695, 1995.

Lattuati, M.: Contribution a l’etude du bilan de l’ozone tro-pospherique a l’interface de l’Europe et de l’Atlantique Nord:3225

modelisation lagrangienne et mesures en altitude, Phd thesis,Universite P.M.Curie, Paris, France, 1997.

LeBras, G., Ravishankara, A., and Fish, D.: Oxidation of at-mospheric reduced sulphur compounds: Perspective from lab-oratory studies - Discussion, Philosophical Transactions of the3230

Royal Society of London Series B-Biological Sciences, 352,181–182, 1997.

Lee, Y. N. and Schwartz, S. E.: Precipitation scavenging, dry de-position and resuspension, vol. 1., chap. Kinetics of oxidationof aqueous sulfur (IV) by nitrogen dioxide, Elsevier, New York,3235

1983.Lightfoot, P. and Cox, R.: Organic peroxy-radicals - kinetics, spec-

troscopy and tropospheric chemistry, Atmospheric Environment,26, 1805–1961, 1992.

Loosmore, G. and Cederwall, R.: Precipitation scavenging of at-3240

mospheric aerosols for emergency response applications: testingan updated model with new real-time data, Atmospheric Envi-ronemnt, 38, 993–1003, 2004.

Louis, J., Tiedke, M., and Geleyn, J.: A short history of the PBLparametrization at ECMWF, in: ECMWF Workshop on Plane-3245

tary Boundary Layer parametrization, pp. 59–80, University ofReading, 1982.

Madronich, S., McKenzie, R. E., Bjorn, L. O., and Caldwell, M. M.:Changes in biologically active ultraviolet radiation reaching theEarth’s surface, J. Photochem. Photobiol. B:Biology, 46, 5–19,3250

1998.Marticorena, B. and Bergametti, G.: Modeling the atmospheric dust

cycle: 1 Design of a soil derived dust production scheme, J. Geo-phys. Res., 100, 16 415–16 430, 1995.

Martinez, R. D., Buitrago, A. A., Howell, N. W., Hearn, C. H., and3255

Joens, J. A.: The near UV absorption-spectra of several aliphatic-aldehydes and ketones at 300-K, Atmospheric Environment, 26,785 – 792, 1992.

McRae, G., Goodin, W., and Seinfeld, J.: Development of a secondgeneration mathematical model for urban air pollution: I. model3260

formulation, Atmospheric Environment, 16, 679–696, 1982.Mentel, T. F., Bleilebens, D., and Wahner, A.: A study of nighttime

nitrogen oxide oxidation in a large reaction chamber - The fateof NO2, N2O5, HNO3, and O3 at different humidities, Atmo-spheric Environment, 30, 4007 – 4020, 1996.3265

Menut, L.: Adjoint modelling for atmospheric pollution processessensitivity at regional scale during the ESQUIF IOP2, J. Geo-phys. Res., 108, 8562, doi:10.1029/2002JD002549, 2003.

Menut, L.: Sensitivity of hourly Saharan dust emissions toNCEP and ECMWF modelled wind speed, Journal of Geo-3270

physical Research, Atmospheres, 113, D16 201, doi:10.1029/2007JD009522, 2008.

Menut, L. and Bessagnet, B.: Atmospheric composition forecastingin Europe, Annales Geophysicae, 28, 61–74, 2010.

Menut, L., Vautard, R., Beekmann, M., and Honore, C.: Sensitiv-3275

ity of Photochemical Pollution using the Adjoint of a Simpli-fied Chemistry-Transport Model, J. Geophys. Res., 105, 15 379–15 402, 2000a.

Menut, L., Vautard, R., Flamant, C., Abonnel, C., Beekmann, M.,Chazette, P., Flamant, P., Gombert, D., Guedalia, D., Lefebvre,3280

M., Lossec, B., D., M., Megie, G., Perros, P., Sicard, M., andToupance, G.: Measurements and modelling of atmospheric pol-lution over the Paris area: an overview of the ESQUIF Project,Annales Geophysicae, 18, 1467–1481, 2000b.

Menut, L., Coll, I., and Cautenet, S.: Impact of meteorological data3285

resolution on the forecasted ozone concentrations during the ES-COMPTE IOP 2a and 2b, Atmospheric Research, 74, 139–159,2005a.

Menut, L., C.Schmechtig, and B.Marticorena: Sensitivity of thesandblasting fluxes calculations to the soil size distribution accu-3290

racy, Journal of Atmospheric and Oceanic Technology, 22, 1875–1884, 2005b.

Menut, L., G.Foret, and G.Bergametti: Sensitivity of mineral dustconcentrations to the model size distribution accuracy, Journalof Geophysical Research, Atmospheres, 112, D10 210, doi:10.3295

1029/2006JD007766, 2007.Menut, L., Chiapello, I., and Moulin, C.: Previsibility of min-

eral dust concentrations: The CHIMERE-DUST forecast duringthe first AMMA experiment dry season, Journal of Geophys-ical Research, Atmospheres, 114, D07 202, doi:doi:10.1029/3300

2008JD010523, 2009a.Menut, L., Masson, O., and Bessagnet, B.: Contribution of

Saharan dust on radionuclides aerosols activity levels in Eu-rope? The 21-22 February 2004 case study, Journal of Geo-physical Research, Atmospheres, 114, D16 202, doi:10.1029/3305

2009JD011767, 2009b.Menut, L., Bessagnet, B., Colette, A., and Khvorostyanov, D.: On

the impact of the vertical resolution on chemistry transport mod-elling, Atmospheric Environment, 67, 370–384, 2013.

Menut, L., Goussebaile, A., Bessagnet, B., Khvorostiyanov, D.,3310

and Ung, A.: Impact of realistic hourly emissions profiles onair pollutants concentrations modelled with CHIMERE, AT-MOSPHERIC ENVIRONMENT, 49, 233–244, doi:{10.1016/j.


atmosenv.2011.11.057}, 2012.Middleton, P., Stockwell, W. R., and Carter, W. P.: Agregation3315

and analysis of volatile organic compound emissions for regionalmodelling, Atmospheric Environment, 24, 1107–1133, 1990.

Milford, J. and Davidson, C.: The sizes of particulate trace elementsin the atmosphere-a review, J. Air Pollution Control Association,35, 1249–1260, 1985.3320

Mircea, M. and Stefan, S.: A theoretical study of the microphysicalparameterization of the scavenging coefficient as a function ofprecipitation type and rate, Atmos. Env., 32, 2931–2938, 1998.

Molnar, A., Meszaros, E., Polyak, K., Borbely-Kiss, I., Koltay, E.,G., S., and Horvath, Z.: Atmospheric budget of different ele-3325

ments in aerosol particles over Hungary, Atmospheric Environ-ment, 29, 1821–1828, 1995.

Monahan, E. C.: In The Role of Air-Sea Exchange in GeochemicalCycling, chap. The ocean as a source of atmospheric particles,pp. 129–163, Kluwer Academic Publishers, Dordrecht, Holland,3330

1986.Monks, P. e. a.: Atmospheric composition change - global and re-

gional air quality, Atmospheric Environment, 43, 5268–5350,2009.

Nenes, A., Pilinis, C., and Pandis, S.: ISORROPIA: A new ther-3335

modynamic model for inorganic multicomponent atmosphericaerosols, Aquatic Geochem., 4, 123–152, 1998.

Odum, J. R., Jungkamp, T. P. W., Griffin, R. J., Forster, H. J. L., Fla-gan, R. C., and Seinfeld, J. H.: Aromatics, reformulated gasolineand atmospheric organic aerosol formation, Environ. Sci. Tech-3340

nol., 31, 1890–1897, 1997.Olivie, D.J.L.van Velthoven, P., Beljaars, A., and Kelder, H.: Com-

parison between archived and off-line diagnosed convective massfluxes in the chemistry transport model TM3, J. Geophys. Res.,109, doi:10.1029/2003JD004036, 2004.3345

Pankow, J. F.: An absorption model of gas/aerosol partition in-volved in the formation of secondary organic aerosol, Atmos. En-viron., 28, 189–193, 1994.

Paulson, S. E. and Seinfeld, J. H.: Development and evaluation of aphotooxidation mechanism for isoprene, Journal of Geophysical3350

Research, 97, 20 703 – 20 715, 1992.Pere, J. C., Mallet, M., Pont, V., and Bessagnet, B.: Impact of

aerosol direct radiative forcing on the radiative budget, surfaceheat fluxes, and atmospheric dynamics during the heat wave ofsummer 2003 over western Europe: A modeling study, J. Geo-3355

phys. Res., 116, D23 119, doi:10.1029/2011JD016240, 2011.Peters, K. and Eiden, R.: Modelling the dry deposition velocity of

aerosol particles to a spruce forest, Atmospheric Environment,26A(14), 2555–2564, 1992.

Pirovano, G., Balzarini, A., Bessagnet, B., Emery, C., Kallos,3360

G., Meleux, F., Mitsakou, C., Nopmongcol, U., Riva, G., andYarwood, G.: Investigating impacts of chemistry and transportmodel formulation on model performance at European scale, At-mospheric Environment, 53, 93 – 109, doi:10.1016/j.atmosenv.2011.12.052, 2012.3365

Pison, I., Menut, L., and N.Blond: Inverse modeling of emissionsfor local photo-oxidant pollution: Testing a new methodologywith kriging constraints, Annales Geophysicae, 24,, 1523–1535,2006.

Pison, I., L.Menut, and G.Bergametti: Inverse modeling of sur-3370

face NOx anthropogenic emissions fluxes in the Paris area duringthe ESQUIF campaign, Journal of Geophysical Research, Atmo-

spheres, 112, D24 302, doi:10.1029/2007JD008871, 2007.Plum, C. N., Sanhueza, E., Atkinson, R., Carter, W. P., and Pitts,

J. N.: Hydroxyl radical rate constants and photolysis rates of3375

alpha-dicarbonyls, Environ. Sci. Technol., 17, 479 – 484, 1983.Priestley, C.: Heat transport and zonal stress between latitudes,

Quaterly Journal of the Royal Meteorology, 75, 28–40, 1949.Pryor, S., Schoof, J., and Barthelmie, R.: Empirical downscaling

of wind speed probability distributions, J. Geophys. Res., 110,3380

D19 109, doi:10.1029/2005JD005899, 2005.Pun, B. K., Seigneur, C., and Lohman, K.: Modeling secondary or-

ganic aerosol formation via multiphase partitioning with molec-ular data, Environ. Sci. Technol., 40, 4722–4731, 2006.

Rao, S., Galmarini, S., and Puckett, K.: Air Quality Model Evalu-3385

ation International Initiative (AQMEII): Advancing the State ofthe Science in Regional Photochemical Modeling and Its Appli-cations, Bull. Amer. Meteor. Soc., 92, 23–30, 2011.

Rio, C., Hourdin, F., and Chedin, A.: Numerical simulation oftropospheric injection of biomass burning products by pyro-3390

thermal plumes, Atmos. Chem. Phys., 10, 3463–3478, doi:10.5194/acp-10-3463-2010, 2010.

Roberts, G. J. and Wooster, M. J.: Fire Detection and Fire Char-acterization Over Africa Using Meteosat SEVIRI, IEEE Trans.Geosci. Remote Sens., 46(4), 1200–1218, 2008.3395

Robinson, A. L., Donahue, N. M., Shrivastava, M. K., Weitkamp,E., Sage, A. M., Grieshop, A. P., Lane, T. E., Pandis, S. N., andPierce, J. R.: Rethinking organic aerosols: Semivolatile emis-sions and photochemical aging, Science, 315, 1259–1262, 2007.

Rogers, C., Wayne, P., Macklin, E., Muilenberg, M., and Wagner,3400

C.: Interaction of the onset of spring and elevated atmosphericCO2 on ragweed (Ambrosia artemisiifolia L.) pollen production,Environ Health Perspect, 114, 865–869, 2006.

Rouıl, L., Honore, C., Vautard, R., Beekmann, M., Bessagnet, B.,Malherbe, L., Meleux, F., Dufour, A., Elichegaray, C., Flaud,3405

J., Menut, L., Martin, D., Peuch, A., Peuch, V., and Poisson,N.: PREV’AIR : an operational forecasting and mapping sys-tem for air quality in Europe, BAMS, 90, 73–83, doi:10.1175/2008BAMS2390.1, 2009.

Royer, P., Chazette, P., Sartelet, K., Zhang, Q. J., Beek-3410

mann, M., and Raut, J.-C.: Comparison of lidar-derived PM10

with regional modeling and ground-based observations in theframe of MEGAPOLI experiment, Atmospheric Chemistry andPhysics, 11, 10 705–10 726, doi:10.5194/acp-11-10705-2011,http://www.atmos-chem-phys.net/11/10705/2011/, 2011.3415

Ryaboshapko, A., Ilyin, I., Gusev, A., Afinogenova, O., Berg, T.,and Hjellbrekke, A.: Monitoring and Modelling of Lead, Cad-mium and Mercury Transboundary Transport in the Atmosphereof Europe, EMEP/MSC-E Report, Meteorological SynthesizingCentre East, Moscow, Russia., p. 1/99, 1999.3420

Rybnicek, O. and Jaeger, S.: Ambrosia (Ragweed) in Europe, in:ACI International, pp. 60–66, Allergy and Clinical ImmunologyInternational, Hofgrefe and Huber Publishers, March/April 2001,Vol 13, No 2, 2001.

Sander, G.C. and. Parlange, J.-Y., Smith, R.E.and Haverkamp, R.,3425

and Hogarth, W.: Estimation of ponding time for constant surfaceflux, Hydrology days Ed. AGU Pub. 19, San Francisco, pp. 402–410, 1999.

Sander, S., Friedl, R., Golden, D., Kurylo, M., Moortgat, G., Keller-Rudek, H., Wine, P., Ravishankara, A., Kolb, C., Molina, M.,3430

Finlayson-Pitts, B., Huie, R., and Orkin, V.: Chemical Kinetics


and Photochemical Data for Use in Atmospheric Studies. Evalu-ation Number 15, JPL Publication 06-2, Jet Propulsion Labora-tory, 2006.

Saunders, S. M., Jenkin, M. E., Derwent, R. G., and Pilling, M. J.:3435

Protocol for the development of the Master Chemical Mech-anism, MCM v3 (Part A): tropospheric degradation of non-aromatic volatile organic compounds, Atmos. Chem. Phys., 3,161–180, 2003.

Schaap, M., Vautard, R., Bergstrom, R., van Loon, M., Bessagnet,3440

B., Brandt, J., Christensen, H., Cuvelier, K., Foltescu, V., Graff,A., E., J. J., Kerschbaumer, A., Krol, M., Langner, J., Roberts,P., Rouil, L., Stern, R., Tarrason, L., Thunis, P., Vignati, E.,White, L., Wind, P., and Builtjes, P. H. J.: Evaluation of long-term aerosol simulations from seven air quality models and their3445

ensemble in the EURODELTA study, Atmospheric Environemnt,41, 2083–2097, 2007.

Schere, K., Flemming, J., Vautard, R., Chemel, C., Colette, A.,Hogrefe, C., Bessagnet, B., Meleux, F., Mathur, R., Roselle,S., Hu, R.-M., Sokhi, R. S., Rao, S. T., and Galmarini, S.:3450

Trace gas/aerosol boundary concentrations and their impactson continental-scale AQMEII modeling domains, AtmosphericEnvironment, 53, 38–50, doi:10.1016/j.atmosenv.2011.09.043,2012.

Schmidt, H., Derognat, C., Vautard, R., and Beekmann, M.: A com-3455

parison of simulated and observed ozone mixing ratios for thesummer of 1998 in western Europe, Atmospheric Environment,35, 6277–6297, 2001.

Seigneur, C. and Constantinou, E.: Chemical Kinetic Mechanismfor Atmospheric Chromium, Environ. Sci. Technol., 29, 222–3460

231, 1995.Seiler, W. and Crutzen, P.: Estimates of gross and net fluxes

of carbon between the biosphere and the atmosphere frombiomass burning, Climatic Change, 2, 207–247, doi:10.1007/BF00137988, 1980.3465

Seinfeld, J. H. and Pandis, S. N.: Atmospheric chemistryand physics: From air pollution to climate change, Wiley-Interscience, J.Wiley, New York, 1997.

Shao, Y. and Lu, I.: A simple expression for wind erosion thresholdfriction velocity, J. Geophys. Res., 105, 22,437–22,443, 2000.3470

Simpson, D.: Long period modeling of photochemical oxidants inEurope, Calculations for July 1985, Atmos. Environ., 26, 1609–1634, 1992.

Skamarock, W., Klemp, J., Dudhia, J., Gill, D., Barker, D., Wang,W., and Powers, J.: A Description of the Advanced Research3475

WRF Version 2, NCAR Technical Note, Boulder, Colorado,USA, pp. NCAR/TN–468+STR, 2007.

Slinn, W.: Precipitation scavenging. Division of Biomedical En-vironmental Research, US Department of Energy, Washington,DC, Atmospheric Sciences and Power Production 1979, 1983.3480

Sofiev, M., Siljamo, P., Ranta, H., and Rantio-Lehtimaki, A.: To-wards numerical forecasting of long-range air transport of birchpollen: theoretical considerations and a feasibility study, Int JBiometeorol, 50, 392–402, 2006.

Sofiev, M., Siljamo, P., Ranta, H., Linkosalo, T., Jaeger, S., Ras-3485

mussen, A., Rantio-lehtimaki, A., Severova, E., and Kukkonen,J.: A numerical model of birch pollen emission and dispersionin the atmosphere. Description of the emission module, Int JBiometeorol, 57-1, i45–58, 2013.

Solazzo, E., Bianconi, R., Pirovano, G., Matthias, V., Vautard, R.,3490

Moran, M. D., Appel, K. W., Bessagnet, B., Brandt, J., Chris-tensen, J. H., Chemel, C., Coll, I., Ferreira, J., Forkel, R., Fran-cis, X. V., Grell, G., Grossi, P., Hansen, A. B., Miranda, A. I.,Nopmongcol, U., Prank, M., Sartelet, K. N., Schaap, M., Silver,J. D., Sokhi, R. S., Vira, J., Werhahn, J., Wolke, R., Yarwood, G.,3495

Zhang, J., Rao, S. T., and Galmarini, S.: Operational model eval-uation for particulate matter in Europe and North America in thecontext of AQMEII, Atmospheric Environment, 53, 75 – 92, doi:10.1016/j.atmosenv.2012.02.045, http://www.sciencedirect.com/science/article/pii/S1352231012001604, 2012a.3500

Solazzo, E., Bianconi, R., Vautard, R., Appel, K. W., Moran,M. D., Hogrefe, C., Bessagnet, B., Brandt, J., Christensen, J. H.,Chemel, C., Coll, I., van der Gon, H. D., Ferreira, J., Forkel,R., Francis, X. V., Grell, G., Grossi, P., Hansen, A. B., Jerice-vic, A., Kraljevic, L., Miranda, A. I., Nopmongcol, U., Pirovano,3505

G., Prank, M., Riccio, A., Sartelet, K. N., Schaap, M., Silver,J. D., Sokhi, R. S., Vira, J., Werhahn, J., Wolke, R., Yarwood,G., Zhang, J., Rao, S., and Galmarini, S.: Model evaluation andensemble modelling of surface-level ozone in Europe and NorthAmerica in the context of AQMEII, Atmospheric Environment,3510

53, 60–74, doi:10.1016/j.atmosenv.2012.01.003, 2012b.Sovde, O. A., Gauss, M., Smyshlyaev, S. P., and Isaksen, I. S. A.:

Evaluation of the chemical transport model Oslo CTM2 with fo-cus on arctic winter ozone depletion, Journal of Geophysical Re-search: Atmospheres, 113, n/a–n/a, doi:10.1029/2007JD009240,3515

http://dx.doi.org/10.1029/2007JD009240, 2008.Stern, R., Builtjes, P., Schaap, M., Timmermans, R., Vautard, R.,

Hodzic, A., Memmesheimer, M., Feldmann, H., Renner, E.,Wolke, R., and Kerschbaumer, A.: A model inter-comparisonstudy focusing on episodes with elevated PM10 concentrations,3520

Atmospheric Environment, 42, 4567–4588, 2008.Stohl, A., Williams, E., Wotawa, G., and Kromp-Kolb, H.: A

European inventory of soil nitric oxide emissions and the ef-fect of these emissions on the photochemical formation ofozone, Atmospheric Environment, 30, 3741 – 3755, doi:10.1016/3525

1352-2310(96)00104-5, 1996.Szopa, S., Foret, G., Menut, L., and Cozic, A.: Impact of large scale

circulation on European summer surface ozone: consequencesfor modeling, Atmospheric Environment, 43, 1189–1195, doi:doi:10.1016/j.atmosenv.2008.10.039, 2009.3530

Thunis, P., Rouil, L., , Stern, R., Kerschbaumer, A., Bessagnet, B.,Builtjes, P., Douros, J., Moussiopoulos, N., Pirovano, G., , andBedogni, M.: Analysis of model responses to emission-reductionscenarios within the City-Delta project, Atmospheric Environ-ment, 41, 208–220, 2007.3535

Tiedtke, M.: A comprehensive mass flux scheme for cumulus pa-rameterization in large-scale models, Monthly Weather Reviews,117, 1779–1800, 1989.

Troen, I. and Mahrt, L.: A simple model of the atmospheric bound-ary layer: Sensitivity to surface evaporation, Bound.-Layer Me-3540

teorol., 37, 129–148, 1986.Tsyro, S.: First estimates of the effect of aerosol dynamics in the

calculation of PM10 and PM2.5, EMEP report, 4/02, NorwegianMeteorological Institute, Oslo, 2002.

Turquety, S.: Fire emission inventory for the Euro-Mediterranean3545

region: variability, uncertainty and adaptability to short-termforecasting, Geosci. Model Dev., in preparation, 2012.

Ullrich, P. A., Jablonowski, C., and Leer, B. V.: High-order finite-volume methods for the shallow-water equations on the sphere, J.


Comput. Phys, 229, 6104–6134, doi:10.1016/j.jcp.2010.04.044,3550

2010.Valari, M. and Menut, L.: Does an increase in air quality mod-

els resolution bring surface ozone concentrations closer to real-ity?, JOURNAL OF ATMOSPHERIC AND OCEANIC TECH-NOLOGY, 25, 1955–1968, doi:{10.1175/2008JTECHA1123.3555

1}, 2008.Valari, M. and Menut, L.: Transferring the heterogeneity of surface

emissions to variability in pollutant concentrations over urbanareas through a chemistry transport model, Atmospheric Envi-ronment, 44, 3229–3238, 2010.3560

Valari, M., Chatignoux, E., and Menut, L.: Using a chemistry trans-port model to account for the spatial variability of exposure-concentrations in epidemiologic air pollution studies, Journal ofthe Air and Waste Management Association, 61, 164–179, 2011.

Van der Gon, D. H., van het Bolscher, M., Visschedijk, A., and3565

Zandveld, P.: Study to the effectiveness of the UNECE HeavyMetals Protocol and costs of possible additional measures. PhaseI: Estimation of emission reduction resulting from the imple-mentation of the HM Protocol, TNO-report, TNO Utrecht, TheNetherlands, p. 193, 2005.3570

van der Werf, G., Randerson, J., Giglio, L., Collatz, G., Mu,M., Kasibhatla, P., Morton, D., DeFries, R., Jin, Y., and vanLeeuwen, T.: Global fire emissions and the contribution of de-forestation, savanna, forest, agricultural, and peat fires (1997-2009), Atmos. Chem. Phys., 10, 11 707–11 735, doi:10.5194/3575

acp-10-11707-2010, 2010.Van Leer, B.: Towards the ultimate conservative difference scheme.

V A second order sequel to Godunov’s method, J. ComputationalPhys., 32, 101–136, 1979.

Van Loon, M., Vautard, R., Schaap, M., Bergstrom, R., Bessagnet,3580

B., Brandt, J., Builtjes, P., Christensen, J. H., Cuvelier, K., Graf,A., Jonson, J., Krol, M., Langner, J., Roberts, P., Rouil, L., Stern,R., Tarrason, L., Thunis, P., Vignati, E., White, L., and Wind, P.:Evaluation of long-term ozone simulations from seven regionalair quality models and their ensemble average, Atmospheric En-3585

vironment, 41, 2083–2097, 2007.Vautard, R., Beekmann, M., and Menut, L.: Applications of ad-

joint modelling in atmospheric chemistry: sensitivity and inversemodelling, Environmental Modelling and Software, 15, 703–709, 2000.3590

Vautard, R., Beekmann, M., Roux, J., and Gombert, D.: Validationof a hybrid forecasting system for the ozone concentrations overthe Paris area, Atmos. Environ., 35, 2449–2461, 2001.

Vautard, R., B.Bessagnet, M.Chin, and Menut, L.: On the contribu-tion of natural Aeolian sources to particulate matter concentra-3595

tions in Europe: testing hypotheses with a modelling approach,Atmospheric Environment, 39, 3291–3303, 2005.

Vautard, R., Maidi, M., Menut, L., Beekmann, M., and Colette,A.: Boundary layer photochemistry simulated with a two-streamconvection scheme, Atmospheric Environment, 41, 8275–8287,3600

2007.Vautard, R., Schaap, M., Bergstrom, R., Bessagnet, B., Brandt, J.,

Builtjes, P., Christensen, J., Cuvelier, C., Foltescu, V., Graf, A.,Kerschbaumer, A., Krol, M., Roberts, P., Rouil, L., Stern, R.,Tarrason, L., Thunis, P., Vignati, E., and Wind, P.: Skill and un-3605

certainty of a regional air quality model ensemble, AtmosphericEnvironment, 39, 2957–2967, 2009.

Vautard, R., Menut, L., Beekmann, M., Chazette, P., Flamant,

P., Gombert, D., Guedalia, D., Kley, D., Lefebvre, M., Mar-tin, D., Megie, G., Perros, P., and Toupance, G.: A syn-3610

thesis of the Air Pollution over the Paris Region (ESQUIF)field campaign, JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES, 108, doi:{10.1029/2003JD003380}, 2003.

Verwer, J.: Gauss-Seidel iteration for stiff ODEs from chemical ki-netics., Journal on Scientific Computing, 15, 1243–1250, 1994.3615

Vestreng, V.: Review and revision of emission data reported toCLRTAP, EMEP Status report, Norwegian Meteorological Insti-tute, Oslo, 2003.

Vestreng, V., Ntziachristos, L., Semb, A., Reis, S., Isaksen, I.S. A., and Tarrason, L.: Evolution of NO(x) emissions in Eu-3620

rope with focus on road transport control measures, Atmo-spheric Chemistry and Physics, 9, 1503–1520, doi:{10.5194/acp-9-1503-2009}, 2009.

Vivanco, M., Palomino, I., Vautard, R., Bessagnet, B., Martin, F.,Menut, L., and Jimenez, S.: Multi-year assessment of photo-3625

chemical air quality simulation over Spain, Env. Mod. and Soft-ware, 24, 63–73, doi:10.1016/j.envsoft.2008.05.004, 2008.

Vivanco, M., Gonzalez, M., Palomino, I., Garrido, J., Querol, X.,Bessagnet, B., de la Rosa, J., and Sanchez de la Campa, A.: Mod-elling Arsenic, Lead, Cadmium and Nickel Ambient Air Concen-3630

trations in Spain, in: Computational Science and Its Applications(ICCSA), 2011 International Conference, Santander Spain, 20-23 june 2011, pp. 243 –246, doi:10.1109/ICCSA.2011.54, 2011.

Vuolo, M., Chepfer, H., Menut, L., and Cezana, G.: Compar-ison of mineral dust layers vertical structures modelled with3635

CHIMERE-DUST and observed with the CALIOP lidar, Jour-nal of Geophysical Research, Atmospheres, 114, D09 214, doi:10.1029/2008JD011219, 2009a.

Vuolo, M., Menut, L., and Chepfer, H.: Impact of transport schemesaccuracy on modelled dust concentrations variability, Journal of3640

Atmospheric and Oceanic Technology, 26, 1135–1143, doi:10.1175/2008JTECHA1197.1, 2009b.

Warren, D. R.: Nucleation and growth of aerosols, Ph.D. the-sis, Dissertation (Ph.D.), California Institute of Technology,Pasadena, 1986.3645

Wesely, M.: Parameterization of Surface Resistances to GaseousDry Deposition in Regional-Scale Numerical Models, Atmo-spheric Environemnt, 23, 1293–1304, 1989.

Wiedinmyer, C., Akagi, S., Yokelson, R., Emmons, L., Al-Saadi, J.,Orlando, J., and Soja, A. J.: The Fire INventory from NCAR3650

(FINN): a high resolution global model to estimate the emis-sions from open burning, Geosci. Model Dev., 4, 625–641, doi:10.5194/gmd-4-625-2011, 2011.

Wilson, R. C., Fleming, Z. L., Monks, P. S., Clain, G., Henne,S., Konovalov, I. B., Szopa, S., and Menut, L.: Have primary3655

emission reduction measures reduced ozone across Europe? Ananalysis of European rural background ozone trends 1996-2005,ACP, 12, 437–454, 2012.

Wu, L., Mallet, V., Bocquet, M., and Sportisse, B.: A comparisonstudy of data assimilation algorithms for ozone forecasts, Jour-3660

nal of Geophysical Research: Atmospheres, 113, doi:10.1029/2008JD009991, http://dx.doi.org/10.1029/2008JD009991, 2008.

Zhang, L., Gong, S., Padro, J., and Barrie, L.: A size-segregatedparticle dry deposition scheme for an atmospheric aerosol mod-ule, Atmos. Environ., 35(3), 549–560, 2001.3665

Zhang, Y., Huang, J.-P., Henze, D. K., and Seinfeld, J. H.:Role of isoprene in secondary organic aerosol formation on


a regional scale, Journal of Geophysical Research: Atmo-spheres, 112, doi:10.1029/2007JD008675, http://dx.doi.org/10.1029/2007JD008675, 2007.3670

Zyryanov, D., Foret, G., Eremenko, M., Beekmann, M., Cam-mas, J.-P., M.D’Isidoro, Elbern, H., Flemming, J., Friese, E.,Kioutsioutkis, I., Maurizi, A., D.Melas, Meleux, F., Menut, L.,Moinat, P., Peuch, V.-H., Poupkou, A., Razinger, M., Schultz,M., Stein, O., Suttie, A. M., Valdebenito, A., Zerefos, C., Du-3675

four, G., Bergametti, G., and Flaud, J.-M.: 3D evaluation of tro-pospheric ozone simulations by an ensemble of regional Chem-istry Transport Model, Atmospheric Chemistry and Physics, 12,3219–3240, 2012.

Documents

CHIMERE 2013: a model for regional atmospheric …menut/pp/2013-gmdd-chimere...Menut L. et al.: CHIMERE: a model for regional atmospheric composition modelling 3 Fig. 1. General principle