tramontane/mistral eventmenut/pp/2006salameh-acpd-0318-ms.pdf · basin in order to provide insight...

This manuscript was prepared for Atmos. Chem. Phys.(Discuss.)with version 7.7 of the egu LATEX class egu.cls.Date: October 11, 2006

ACP-LOGO

Aerosol distribution over the westernMediterranean basin during aTramontane/Mistral event

T. Salameh1, P. Drobinski2, L. Menut1, B. Bessagnet3, C. Flamant2, A. Hodzic4, andR. Vautard5

1Institut Pierre Simon Laplace/Laboratoire de Meteorologie Dynamique, EcolePolytechnique/ENS/UPMC/CNRS, Palaiseau, France2Institut Pierre Simon Laplace/Service d’Aeronomie, UPMC/UVSQ/CNRS, Paris, France3Institut National de l’Environnement Industriel et des Risques, INERIS, Verneuil enHalatte, France4National Center for Atmospheric Research, Boulder, Colorado, USA5Institut Pierre Simon Laplace/Laboratoire des Sciences du Climat et de l’Environnement,CEA/UVSQ/CNRS, Gif sur Yvette, France

Abstract. This paper investigates ex-perimentally and numerically the timeevolution of the spatial distribution ofaerosols over the Western Mediterraneanbasin during March, 24 1998 Mistralevent documented during the FETCHexperiment. Mistral and Tramontaneare very frequent northerly winds (5-15days per month) accelerated along theRhone and Aude valley (France) thatcan transport natural and anthropogenicaerosols offshore as far as a few hundredsof kilometers which can in turn have animpact on the radiation budget over theMediterranean Sea and on precipitation.The spatial distribution of aerosols wasdocumented by means of the airbornelidar LEANDRE-2 and spaceborne ra-diometer SeaWIFS, and a validatedmesoscale chemical simulation using thechemistry-transport model CHIMEREwith an aerosol module, forced by thenon-hydrostatic model MM5.This study shows that: (1) the Mistralcontributes to the offshore exportationof a large amount of aerosols originallyemitted over continental Europe (in partic-ular ammonium nitrate in the particulatephase and sulfates) and along the shorefrom the industrialized and urban areas ofFos-Berre/Marseille. The Genoa surfacelow contributes to advect the aerosolsalong a cyclonic trajectory that skirts the

Correspondence to: T. Salameh([email protected])

North African coast and reaches Italy;(2) the aerosol concentration pattern isvery unsteady as a result of the timeevolution of the two winds (or Genoacyclone position): The Tramontane windprevails in the morning hours of March 24,leaving room for the Mistral wind and anunusually strong Ligurian outflow in theafternoon. The wakes trailing downstreamthe Massif Central and the Alps preventany horizontal diffusion of the aerosolsand can, at times, contribute to aerosolstagnation.

1 Introduction

The Mediterranean basin is featured byan almost-closed ocean basin surroundedby mountain ranges in which numerousrivers rise, a contrasted climate and vege-tation from south to north, numerous andrapidly growing built-up areas along thecoast with several major cities having in-dustrial activities emitting a large numberof gas substances and aerosols. Highlyaerosol loaded air masses are found fromthe surface up to the upper troposphere andcontribute to decrease air quality on a largescale and reduce precipitation in the re-gion (Lelieveld et al. 2002). The aerosolsfound below 4 km height originate from re-gional sources whereas above 4 km heightthey are usually linked to transport due toglobal-scale motions and teleconnections,for instance with the Indian monsoon and

© 2006 Author(s). This work is licensed under a Creative Commons License.

2 T. Salameh et al.: Aerosols distribution during a Tramontane/Mistral episode

the North Atlantic Oscillation (Moulin etal. 1997). High aerosol loads in the midtroposphere may also originate from wind-blown dust in the Saharan regions.

Near the surface, most of the aerosolsoriginate from western and eastern Europe(e.g. Sciare et al. 2003; Traub et al. 2003;Schneider et al. 2004), and from the Sa-haran desert (e.g. Bergametti et al. 1992;Moulin et al. 1998; Guieu et al. 2002).Industrial activity, traffic, forest fires, agri-cultural and domestic burning are the mainsource of pollution in Europe whereas theclose vicinity of the Saharan desert pro-vides a source for considerable amounts ofdust.

Aerosols are harmful for ecosystems andhuman health, and they affect the Mediter-ranean climate and water cycle. Indeed,the microscopic particles affect the atmo-spheric energy budget by scattering and ab-sorbing solar radiation, reducing solar ra-diation absorption by the sea by approxi-mately 10 % and altering the heating pro-file of the lower troposphere (Lelieveld etal. 2002). As a consequence, evapora-tion and moisture transport, in particularto North Africa and the Middle East, aresuppressed. This aerosol effect is substan-tial today, although the period with highestaerosol concentrations over the Mediter-ranean Sea was around 1980 and con-tributed to the drought in the eastern Sahel(Lelieveld et al. 2002). Aerosols also af-fect the Mediterranean biogeochemistry bydeposition (wet and dry) of dissolved inor-ganic phosphorus (Bergametti et al. 1992;Bartoli et al. 2005), silicon (Bartoli etal. 2005) and iron (Guieu et al. 2002)from soil-derived dust from desert areas ofnorth Africa and anthropogenic emissionsfrom European countries. Apart from in-ternal sources of nutrients, the atmosphericinputs may constitute an important path-way for nutrients to the photic zone of theopen sea where there is little riverine in-put (Migon et al. 1989; Bergametti etal. 1992; Prospero et al. 1996; Benitez-Nelson 2000). This particularly applies tothe Mediterranean Sea because of its re-duced dimensions and because surround-ing continental emission sources of nutri-ents are intense and continuously increas-ing (Bethoux 1989; Guerzoni et al. 1999;

Bethoux et al. 2002).In the context of climate change in

which the Mediterranean basin appearsquite vulnerable due to hydric stress andever increasing pollution levels, it is cru-cial that knowledge be improved concern-ing the mechanisms linking the dynamicsof the main flow regimes and the exist-ing pollution sources scattered around thebasin in order to provide insight into futuretrends at the regional scale.

The western Mediterranean climate isfrequently affected by the Mistral andits companion wind, the Tramontane(Georgelin and Richard 1996; Drobinskiet al. 2001a). The Mistral is a se-vere northerly wind that develops along theRhone valley (while the Tramontane blowsin the Aude valley) (Fig. 1), and occurs be-tween 5 and 15 days per month. The devel-opment of the Mistral is preconditioned bycyclogenesis over the Gulf of Genoa andthe passage of a trough through France.The Mistral occurs all year long but ex-hibits a seasonal variability either in termsof its strength and direction, or in termsof its spatial distribution (Mayencon 1982;Orieux and Pouget 1984). At the regionalscale, the Mistral is frequently observed toextend as far as a few hundreds of kilome-tres from the coast (Jansa 1987) and is thusassociated with low continental pollutionlevels near the coastline as it advects thepollutants away from their sources of emis-sion over the Mediterranean Sea (Bastin etal. 2006; Drobinski et al. 2006).

Due to favourable dispersion conditions,air quality studies generally do not focuson Mistral events. However the transportand resulting concentration distribution ofchemical compounds and aerosols over theMediterranean Sea has never been docu-mented. The FETCH (Flux, Etat de mer etTeledetection en Condition de fetcH vari-able) experiment (Hauser et al. 2003) of-fers an ideal framework for such a study.The FETCH experiment took place fromMarch 12 to April 15, 1998 and was dedi-cated to improve the knowledge of the in-teractions between the ocean and the at-mosphere in a coastal environment understrong wind conditions (e.g. Drennan etal. 2003; Eymard et al. 2003; Flamantet al. 2003). The March 24, 1998 Mis-

Atmos. Chem. Phys., 0000, 0001–25, 2006 www.atmos-chem-phys.org/acp/0000/0001/

T. Salameh et al.: Aerosols distribution during a Tramontane/Mistral episode 3

−5 0 5 10 1536

40

44

48

52

Longitude (°E)

Latit

ude

(°N

)

(a)

MassifCentral

Alps

Pyrénées

FRANCE

AtlanticOcean

Mediterranean Sea

ITALYAppenine range

TyrrhenianSea

LigurianSea

1 2 3 4 5 6 7 8 9 1039

40

41

42

43

44

45

Longitude (°E)

Latit

ude

(°N

)

(b)LYO

NIM

VAL

TLN

Rhô

ne v

alle

y

Aude valley

GEN

MILTOR

BAR

Gulfof Lion

Corsica

Sardinia

BalearicIslands

3.43.94.44.95.441.8

42.2

42.6

43

43.4

43.8

44.2

Longitude (°E)

Latit

ude

(°N

)

(c)NIM

MRS

AIXFOS

MONA

F

E

C

Fig. 1. Panel a: Map of France with the topography shaded in grey when higher than 500 m abovesea level. The two rectangles in dashed line display the large and medium domains (hereafter calleddomain 1 and domain 2, respectively) of the MM5 simulations. Panel b: Domain 2 of the MM5simulation with its nested smaller domain (hereafter called domain 3) in the rectangle in dashed line.The acronyms BAR, GEN, MIL, NIM, LYO, TLN, TLS, TOR and VAL stand for the city namesBarcelona, Genoa, Milano, Nimes, Lyon, Toulon, Toulouse, Torino and Valence, respectively. Thedots indicate the locations of the operational meteorological surface stations. Panel c: Zoom onthe Gulf du Lyon area. The rectangle in dashed line displays the small domain (domain 3) of theMM5 simulations. The dotted line corresponds to the flight track of the ARAT carrying the watervapor lidar LEANDRE-2 on March 24, 1998 in the afternoon and the thick solid line to the route ofR/V ATALANTE. The diamond marker shows the location of the ASIS buoy. The acronyms AIX,FOS, MON, MRS and NIM correspond to the city names Aix en Provence, Fos/Berre, Montpellier,Marseille and N ımes, respectively.

tral case is a strong episode, typical ofintermediate season Mistral events com-pared to weaker summer Mistral events(e.g. Drobinski et al. 2005, Bastin etal. 2006). Flamant (2003) analyzed thecomplex structure of atmospheric bound-ary layer (ABL) observed using an air-borne lidar over the Gulf of Lion at theexit of the Rhone valley during the March,24 1998 Mistral event. Even though thestudy by Flamant (2003) suggested the ex-istence of a marked east-west aerosol con-centration gradient offshore to be related tolarger concentrations of pollution aerosol

from the city of Marseille and the indus-trial petrochemical complex of Fos/Berre,the highly spatially resolved aerosol mea-surements needed to (un)validate this hy-pothesis simply did not exist. By combin-ing lidar observations and simulations webetter address this question in the presentarticle. Therefore this article is designed toanalyze:

– the dynamic processes driving thesmall-scale structure of the Mistralflow at the exit of the Rhone valley,and their relation with the aerosol dis-

www.atmos-chem-phys.org/acp/0000/0001/ Atmos. Chem. Phys., 0000, 0001–25, 2006


tribution observed by the airborne li-dar and the satellite imagery,

– the aerosol sources, composition anddistribution over the whole westernMediterranean basin.

The instrument set-up (i.e. FETCH-related sea-borne, airborne and space-borne observations) and the numericalmodels used in this study are describedin Section 2. In Section 3, the meteoro-logical environment leading to the Mistralepisode is analyzed as well as the fine-scalestructure of the Mistral flow. In Section 4,the aerosol distribution over the westernMediterranean basin is discussed, beforeconclusions are drawn in Section 5 andsuggestions for future work are presented.

2 Measurements and models

2.1 Numerical models

2.1.1 Dynamical model

The dynamical numerical simulationspresented in this study have been con-ducted using the fifth generation PennState-National Center for Atmospheric Re-search MM5 model, version 3.6 (Dudhia1993; Grell et al. 1995). The model solvesthe non-hydrostatic equations of motionin a terrain-following sigma coordinates.Three interactively nested model domainsare used, the horizontal mesh size being27 km, 9 km and 3 km, respectively. Do-mains 1 (coarse domain), 2 (medium do-main) and 3 (fine domain) are centeredat 43.7◦N, 4.6◦E and cover an area of1350 km × 1350 km, 738 km × 738 kmand 120 km × 174 km, respectively (seeFig. 1). Domain 1 covers half of Franceand the western Mediterranean Sea. Do-main 2 covers the Rhone valley, the west-ern Alps, the Massif Central and Cor-sica. Domain 3 covers the Rhone valleydelta. The model orography of the threedomains is interpolated from terrain datawith 30” resolution. It is filtered by a two-pass smoother-desmoother (Guo and Chen1994) in order to remove two-grid intervalwaves that would induce numerical noise.Information on land use was obtained from

USGS (United States Geological Survey)data with the same horizontal resolution asfor orography. In the vertical, 43 unevenlyspaced full sigma-levels are used, corre-sponding to 42 half-sigma levels wheremost of the variables are computed. Thelowermost half-sigma level (σ = 0.999) isabout 12 m above ground. The verticaldistance between the model levels is about50 m close to the ground and increases upto 1200 m near the upper boundary whichis located at 100 hPa.

A complete set of physics parameter-izations is used. Cloud microphysics istreated with a sophisticated scheme hav-ing prognostic equations for cloud water,cloud ice, cloud ice particle number con-centration, rain, snow and graupel (Reisneret al. 1998), using modifications proposedby Chiriacco et al. (2005). The Grell cu-mulus parametrization (Grell 1993) is usedin model domains 1 and 2. In model do-main 3, a cumulus parametrization is notneeded because convection is resolved ex-plicitly at such high resolution. The ra-diation scheme accounts for the interac-tion with moisture and clouds (Grell et al.1995; Mlawer et al. 1997). The ABLis parameterized using the Hong and Panscheme (Hong and Pan 1996). It is an ef-ficient scheme based on Troen and Mahrtrepresentation of countergradient term andeddy viscosity profile in the well mixedABL, and is suitable for high resolution inABL.

The initial and boundary conditions aretaken from the ECMWF (European Cen-tre for Medium Range Weather Forecast)reanalyses ERA-40. These reanalysis dataare available every six hours on a 1◦

× 1◦

latitude-longitude grid. Since the interpo-lation routine of the MM5 modeling sys-tem needs pressure level data, the standard-level-pressure version of the ECMWF datais used. In order for the synoptic flowto stay close to meteorological analyzes,the meteorological simulation uses nudg-ing for all variables with a coefficient of10−4 s−1. The initialization date is March23, 1998 at 1200 UTC and the simulationends on 25 March 25, 1998 at 1800 UTC.

In the following, MM5 simulations areused to provide the three-dimensional en-vironment necessary for the interpreta-



tion of the measurements, and are alsothe meteorological forcing for the regionalchemistry-transport model CHIMERE (seenext section).

2.1.2 Chemistry-transport model

The chemistry-transport model forcedby the MM5 meteorology is CHIMERE.It is a 3D chemistry-transport model, de-veloped cooperatively by the French In-stitute Pierre-Simon Laplace (IPSL), Lab-oratoire Interuniversitaire des SystemesAtmospheriques (LISA) and Institut Na-tional de l’Environnement industriel et desRISques (INERIS). CHIMERE is a 3-Dchemistry transport model that simulatesgas-phase and aerosol-phase chemistry andits transport at the scale of the conti-nent (Schmidt et al. 2001) or at regionalscale (Vautard et al. 2001; Vautard et al.2003). It simulates the concentration of44 gaseous species and 6 aerosol chemicalcompounds. Sea salt is not accounted forin our simulations. In CHIMERE aerosolversion (Bessagnet et al. 2004; Hodzic etal. 2004), the population of aerosol par-ticles is represented using a sectional for-mulation, assuming discrete aerosol sizesections and considering the particles of agiven section as homogeneous in composi-tion (internally mixed). Like in the work ofBessagnet et al. (2004), we use six diame-ter bins ranging between 10 nm and 40 µm,with a geometric increase of bin bounds.The aerosol module accounts for both in-organic and organic species, of primary orsecondary origin, such as, primary partic-ulate matter (PPM), sulfates, nitrates, am-monium, secondary organic species (SOA)and water. PPM is composed of primaryanthropogenic species such as elementaland organic carbon and crustal materials.In the model, ammonia, nitrate, and sulfateare considered in aqueous, gaseous, andparticulate phases.

In the present application, the modelis run at mesoscale over a domain cov-ering the western Mediterranean basin(Fig. 1). The model domains matchthe MM5 domains with a vertical resolu-tion of 12 sigma-pressure levels extend-ing up to 500 hPa that covers the ABLand the lower part of the free troposphere.

The meteorological variables used as in-put are wind, temperature, mixing ratio forwater vapor and liquid water in clouds,2 m temperature, surface heat and mois-ture fluxes and precipitation. Boundaryconditions of CHIMERE simulations aretaken from climatologies of LMDZ-INCAchemistry transport model for the gazeousspecies, and from GOCART (The Geor-gia Tech/Goddard Global Ozone Chem-istry Aerosol Radiation Transport) aerosolforecasting model for the aerosols. Emis-sions of primary particulate matter PM10and PM2.5 (particulate matter of diam-eter less than 10 µm and 2.5 µm) andthe anthropogenic emissions for NOx, CO,SO2, NMVOC (Non Methane Volatile Or-ganic Compounds) and NH3 gas-phasespecies for 10 anthropogenic activity sec-tors (as defined by the SNAP cate-gories) are provided by EMEP (availableat http://www.emep.int) with spatial reso-lution of 50 km (Vestreng 2003).

The CHIMERE simulation covers thesame period as the MM5 simulation. How-ever, for aerosol-related spin-up needs, athree-day simulation was ran from March20 1200 UTC to March 23 1200 UTC toaccount for some of the aerosol distributionbuild-up prior to the event under scrutiny.

2.2 Observations

Over the continent, we mainly use themeasurements from the operational ra-diosoundings and synoptic meteorologi-cal stations of Meteo-France. On March24, 1998, operational radiosondes werereleased every twelve hours from Lyonand Nımes. At the surface, the opera-tional meteorological surface station net-work of Meteo France gave access to thesurface thermodynamical fields (precipita-tion, wind speed and direction, tempera-ture, humidity, pressure). The locationsof the operational meteorological surfacestations used in this study are shown inFig. 1b.

Over the Mediterranean Sea, we usethe measurements of meteorological andaerosols variables collected on board theresearch vessel (R/V) Atalante, the ASIS(Air-Sea Interaction Spar) buoy and theAvion de Recherche Atmospherique et



Teledetection (ARAT), as well as fromsatellites (the Sea-viewing Wide Field-of-view Sensor -SeaWIFS- and the ActiveMicrowave Instrumentation-Wind Scat-terometer AMI/Wind on board the Euro-pean Remote Sensing Satellite ERS). Thecomplementarity of the data provided bythese various instruments is an essential as-pect of this study. The locations of themeasurement sites used in this study areshown in Fig. 1.

The R/V Atalante had balloon launch-ing capability and carried an instrumentedmast for mean and turbulent measurementsat a height of 17 m above the sea surface.The ASIS buoy made measurements ofmean and turbulent atmospheric variables7 m above the air-sea interface as well aswave directional spectra. The ARAT wasequipped with standard in situ sensors aswell as sensors dedicated to the analysisof aerosol properties (nephelometer, par-ticle and cloud condensation nuclei coun-ters). The ARAT also embarked the differ-ential absorption lidar LEANDRE-2 (Fla-mant, 2003; Flamant et al., 2003).

The ARAT afternoon mission on March24, 1998 (between 1620 and 1850 UTC)was designed in such a manner that longlevelled legs be flown along and across themean ABL wind direction. The structureof the flow and its evolution with time hasbeen analyzed using high resolution mea-surements (15 m in the vertical and 160 min the horizontal) of atmospheric reflectiv-ity at 730 nm, made with the airborne lidarLEANDRE-2. Because it is sensitive torelative humidity as well as aerosol prop-erties and concentration, lidar-derived re-flectivity is extremely useful to investigateABL structural properties (e.g. Flamantand Pelon 1996; Drobinski et al. 2001a;Flamant 2003). Here we shall discuss re-flectivity measurements made with a nadirpointing system from an altitude of 4 kmabove mean sea level (MSL).

The high spatial and temporal documen-tation of the 150 km × 200 km FETCHtarget area (i.e. Gulf of Lion) allows avery detailed insight on the dynamics ofthe Mistral event as well as on the aerosoldistribution. However, this dataset wascomplemented by satellite data from AMI-Wind/ERS scatterometer providing daily

mean surface wind speed and wind stressover the whole Mediterranean Sea at 1◦

resolution, and from SeaWIFS providingthe aerosol optical depth at 865 nm whilepassing over the western Mediterranean at1132 UTC.

3 The March 24, 1998 Mistral event

3.1 Synoptic environment

The March 24, 1998 Mistral event wasfeatured by the existence of an upper leveltrough associated with a cold front pro-gressing toward the Alps and a shallowvortex (1014 hPa) over the Tyrrhenian Seabetween Sardinia and continental Italy, at0600 UTC. As the day progressed, the lowover the Tyrrhenian deepened (from 1014to 1008 hPa between 0600 and 1500 UTC)while remaining relatively still. From1500 UTC on, the low continued to deepen(from 1008 to 1002 hPa) while moving tothe southeast. It was located over Sicily onMarch 25, 1998 at 0600 UTC.

This two-phase evolution of Alpine leecyclogenesis has been observed before(Egger, 1972; Buzzi and Tibaldi, 1978).Alpert et al. (1996) have shown that to-pographical blocking is the dominant fac-tor in the first and most rapid phase ofthe cyclone deepening. During the secondphase, the growth rate drops to baroclinicvalues and the structure of the cyclone ap-proaches that of typical extratropical sys-tems. Convection as well as sensible andlatent heat fluxes play an important rolein the second phase of Alpine lee cyclo-genesis development (Alpert et al., 1996;Grotjahn and Wang, 1989). These authorshave also shown that convection has a ten-dency to move the cyclone to the eastnorth-east while topography has a tendency to tiethe cyclone to the lee of the Alps. The seamoisture fluxes tend to move the cyclonetoward the warm bodies of water (i.e., to-ward Sicily in our case).

The multistage evolution of the Alpinelee cyclone over the Tyrrhenian Sea in-duced a very nonstationary wind regimeover the Gulf of Lion (also see Flamant,2003). The diurnal evolution of the Mis-tral and the Tramontane on March 24, 1998



are evidenced on the wind field simulatedin the ABL (at 950 hPa) by the MM5model at 0600, 0900, 1200, 1500, 1800 and2100 UTC (Fig. 2). In the early stage (lowat 1014 hPa, 0600 UTC), the Tramontaneflow prevailed over the Gulf of Lion. Thelarge westerly flow component (leading toprevailing Tramontane conditions) resultedfrom the rather high position (in terms oflatitude) of the depression. The Mistralextended all the way to Southern Corsica,wrapping around the depression. To thenorth, a weak easterly outflow was ob-served over the Gulf of Genoa. As the lowdeepened (1010 hPa), the prevailing windregime shifted to a well-established Mistralpeaking around 1200 UTC. The Mistralwas observed to reach Southern Sardiniawhere it wrapped around the depression.At this time, the outflow from the Lig-urian Sea (i.e., Gulf of Genoa) had becomestronger. In the afternoon, the Mistral wasprogressively disrupted by the strengthen-ing outflow coming from the Ligurian Seain response to the deepening low over theTyrrhenian Sea (1008 hPa, 1500 UTC) andthe channelling induced by the presence ofthe Apennine range (Italy) and the Alps. Inthe evening, the Mistral was again well es-tablished over the Gulf of Lion as the de-pression continued to deepen (1002 hPa,2100 UTC), but moved to the southeast re-ducing the influence of outflow from theLigurian Sea on the flow over the Gulf ofLion. During this period, the Tramontaneflow appeared to be much steadier than theMistral and less disrupted by the returnflow associated with the depression. Thecold air outbreak episode over the Gulf ofLion ended on March 25 at 0600 UTC andanticyclonic conditions then prevailed overthe Gulf of Lion.

An important feature of the cold air out-break over the Gulf of Lion is also ob-served in the form of banners of weakerwinds (sheltered region) separating (i) theMistral and the Tramontane (in the lee ofthe Massif Central) and (ii) the Mistraland the Ligurian Sea outflow (in the lee ofthe western Alps, see Fig. 2). Such shel-tered regions are a common feature overthe Mediterranean because of the presenceof numerous mountain ranges and are re-lated to potential vorticity banners (i.e., re-

gions of important wind shear) generateddownstream of the mountain ranges (e.g.Aebischer and Schar 1998; Drobinski et al.2005; Guenard et al. 2006), which con-tribute to the deepening of the Alpine leecyclones. The unsteady nature of the west-ern Alps wake (as opposed to the stead-ier Massif Central wake) is caused by thecomplex topography of the Alps which am-plifies any variation of the wind speed anddirection, or ABL depth upstream of theAlps (Guenard et al. 2006).

3.2 Structure of the Mistral over the con-tinent

The radiosoundings launched from Lyonon March 24, 1998 at 0000 and 1200 UTCshow the synoptic northwesterly flowblowing above 1 km ASL (Fig. 3, top row).Below 1 km height, a low-level jet blowsat about 5-8 m s−1 from the north and isaligned with the Rhone valley axis display-ing channelling in the valley. Between 1and 3 km ASL, the wind direction veers tothe northeast and the wind speed increasesfrom about 5 m s−1 to about 15 m s−1 at3 km ASL. The Mistral jet blows within theABL. The radiosoundings launched fromNımes (Fig. 3, middle row) shows thatthe Mistral experiences channelling in theRhone valley and accelerates with the windspeed increasing from 5-8 m s−1 in Lyonto 15-18 m s−1 in Nımes. The daytimecooling between the entrance and exit ofthe Rhone valley is reversed during night-time because of the large continental diur-nal cycle in Lyon and the smoothed diur-nal cycle near the sea shore. The MM5model (solid line) accurately reproducesthe vertical profiles of wind speed and di-rection, and potential temperature with atmost 2 m s−1 difference, less than 20◦ and1 K on average, respectively. Except fornear surface temperature, the radiosound-ings launched from Lyon and Nımes onMarch 24, 1998 at 0000 and 1200 UTCshow that the Mistral characteristics arefairly persistent with time.

The simulated surface wind and tem-perature fields are evaluated by comparingthe measured 10-m horizontal wind com-ponents and 2-m temperature to their sim-ulated counterparts interpolated at the loca-



Fig. 2. Horizontal wind fields simulated with MM5 (at 950 hPa) and extracted from domain 2 onMarch 24, 1998 at 0600 (a), 0900 (b), 1200 (c), 1500 (d), 1800 (e) and 2100 UTC (f), respectively.The isocontours represent the isotachs. Contour interval is 2 m s−1 from 10 to 20 m s−1. The blacksolid line represent the coastline.

tion of the meteorological surface stationsin southern France (see dots in Fig. 1b).Quantitatively, Fig. 4 shows the histogramsof the difference between the three-hourly

simulated and measured 10-m zonal andmeridional wind components and 2-m tem-perature during the 24 hours on March 24,1998. The measured 10-m wind speed and



3 13 23 33 43012345

Hei

ght A

SL (

km)

(a)

MM5 0000 UTCObs. 0000 UTCMM5 1200 UTCObs. 1200 UTC

180 270 0 90 180012345

(b)

274 284 294 304 314012345

(c)

3 13 23 33 43012345

Hei

ght A

SL (

km)

(d)

180 270 0 90 180012345

(e)

274 284 294 304 314012345

(f)

3 13 23 33 43012345

Wind speed (m s−1)

Hei

ght A

SL (

km)

(g)

180 270 0 90 180012345

Wind direction (o)

(h)

274 284 294 304 314012345

Potential temperature (oK)

(i)

Fig. 3. Vertical profiles of wind speed, wind direction and potential temperature (left, middle and rightcolumns, respectively) over Lyon, N ımes and on the location of the R/V ATALANTE (top, middleand bottom rows, respectively). The measured and simulated profiles are displayed with dashedand solid lines, respectively. On the first and the second rows, thin lines represent measurements at0000 UTC and bold lines represent measurements at 1200 UTC. On the third row, thin lines are formeasurements at 0900 UTC and bold lines are for measurements at 1100 UTC. The location of theR/V ATALANTE was 42.75◦N/4.37◦E at 0900 UTC and 42.95◦N/4.27◦E at 1100 UTC.

2-m temperature accuracies are 1 m s−1

and about 0.1◦ K, respectively. The biasbetween MM5 and the measured windspeed is about -1 m s−1 for the two hori-zontal components which is thus non sig-nificant. The standard deviation is alsosmall considering the accuracy of the windobservations. As for the temperature, thelargest discrepancies are found in the steeporography regions where the height of thetopography is not accurately representedin the model and at night when numericaldiffusion slightly deteriorates the simula-tion of the near surface temperature and thekatabatic flows.

3.3 Structure of the Mistral over theMediterranean Sea

Over the Mediterranean Sea, the dataare generally sparse, and available fromsatellite-borne sensors thus restricted to the

surface and reliable few hundreds of kilo-meters away from the shore. Figure 5shows the daily mean values of the sea sur-face wind speed provided by the satellite-borne AMI-Wind/ERS scatterometer overthe Mediterranean Sea at 1◦ resolution (nowind vectors are available for year 1998).It shows three regions of sea surface windspeed exceeding 20 m s−1: to the west,the Tramontane merge with the SpanishCierzo; at around 7◦E longitude, the Mis-tral blows around the Genoa cyclone andimmediately to the east of Sicily and Sar-dinia, the Ligurian outflow. The MM5models reproduces the same structure butwith lower wind speed which can reach5 m s−1 in the regions of strong flows.From an observational point of view, thiscan be attributed to the fact that the largestAMI-Wind/ERS errors are found near theshore, whereas from a modelling point ofview, the absence of coupling between the



−20−15−10 −5 0 5 10 15 200

100

200

300

# da

ta

u10 mmodel−u

10 msurface (m s−1)

Bias = −1.0387 Std = 3.1059

−20−15−10 −5 0 5 10 15 200

100

200

300

# da

ta

v10 mmodel−v

10 msurface (m s−1)

Bias = 0.0997 Std = 3.5512

−20−15−10 −5 0 5 10 15 200

100

200

300

400#

data

T2 mmodel−T

2 msurface (°K)

Bias= −2.1713 Std= 2.8951

Fig. 4. Histograms of the difference between the three-hourly simulated and measured 10-m zonal (a)and meridional (b) wind components and 2-m temperature (c) during the 24 hours period of March 24,1998 in southern France (see dots in Fig. 1b). The measured 10-m wind speed and 2-m temperatureaccuracies are 1 m s−1 and about 0.1◦ K, respectively.

sea and the atmosphere for winter Mistralevents which generate at this period of theyear intense air-sea heat exchanges (Fla-mant 2003) and sea surface cooling (Mil-lot 1979) can also be a source of error.The ASIS buoy data allows a more detailedcomparison at a single point. Figure 6shows the good agreement between thesurface wind speed and direction, temper-ature, relative humidity and surface stressfrom the ASIS buoy measurements andfrom the MM5 simulations (better than2 m s−1, 10◦, 1◦C, 20 % and 0.1 N m−2

for wind speed and direction, temperature,relative humidity and wind stress, respec-tively, with some intermittent exceptions).This figure illustrates the unsteady aspectof the Mistral/Tramontane episode withthe Tramontane blowing over the buoybetween 0400 and 0800 UTC, followedby a calm period until the Mistral breaksthrough after 1200 UTC.

The radiosoundings launched at0900 UTC and 1100 UTC from the re-search vessel Atalante at (42.75◦N;4.37◦E

longitude) and (42.95◦N;4.27◦E), respec-tively show a northerly wind flow from thesurface up to 5 km and more (Fig. 3g, hand i). The vertical profiles of potentialtemperature at 0900 and 1100 UTC showa remarkable persistence of the thermo-dynamical characteristics of the Mistralwith a potential temperature inversion atabout 2 km height which corresponds tothe ABL depth. At 0900 UTC, the windblows at approximately 15 m s−1 withinthe ABL and increases up to 28 m s−1

above (Fig. 3g). The wind speed isstronger at 1100 UTC than at 0900 UTC(20-22 m s−1 below 3 km) since the re-search vessel Atalante moves closer to theMistral core. The MM5 model reproducesthe vertical profiles of wind speed anddirection, and potential temperature withat most 2 m s−1 (0900 UTC) to 5 m s−1

(1100 UTC) difference and less than 30◦

and 2 K on average for the wind speedand direction and potential temperature,respectively. The deepening of the ABLbetween Nımes and the Atalante can be



−1 1 3 5 7 9 11 1337

39

41

43

45

47

NSCAT

Longitude (oE)

Lat

itude

(o N

)(a)

Win

d sp

eed

(m s

−1 )

0

5

10

15

20

−1 1 3 5 7 9 11 1337

39

41

43

45

47

MM5

Longitude (oE)

Lat

itude

(o N

)

(b)

−1 1 3 5 7 9 11 1337

39

41

43

45

47

Longitude (oE)

Lat

itude

(o N

)

(c)

Win

d sp

eed

diff

eren

ce (

m s

−1 )

−5

−4

−3

−2

−1

0

1

Fig. 5. (a), (b) and (c) are the daily mean values of the wind speeds measured on March 24, 1998by the AMI-Wind/ERS satellite over sea surface, simulated by MM5 (domain 3) and the differencebetween the measured and simulated wind speeds, respectively. (a) and (b) are with the same colorscale.

attributed to the occurrence of a hydraulicjump associated with enhanced turbulenceand mixing (e.g. Drobinski et al. 2001b)as previously observed at the exit of theRhone valley for many other Mistralevents by Pettre (1982), Corsmeier et al.(2005) and Drobinski et al. (2005).

4 Aerosol distribution at the scale ofthe western Mediterranean basin

4.1 Sources of atmospheric pollutants

The western Mediterranean basin ischaracterized by a meridional gradient ofsurface land types, mainly associated withvegetation and aerosol sources. South ofthe basin, the desert is a primary sourceof mineral dust. West and north of thebasin, some aerosol sources are large in-dustrial zones (refineries, power plants,...)surrounded by residential areas with im-

portant traffic, such as Barcelona, Mar-seille, Lyon, Milano and Rome. Ruralareas prone to agriculture or covered byMediterranean natural landscape are themain sources of biogenic emissions. Fi-nally emissions related to ship traffic alsoimpact Mediterranean pollution levels andclimate (Marmer and Langmann 2005).

Some anthropogenic species such as ni-trogen oxides (Fig. 7 a) are mainly emittedvia traffic, industries and domestic com-bustion. For instance, over the continent,maxima of NO emissions are seen in themain cities Toulouse, Lyon, Marseille aswell as Milano. Over the Sea, evidenceof emissions associated with maritime traf-fic is also seen along the northern Africancoastline (i.e. ship routes between Gibral-tar and the Suez canal), as well as along theSpanish coastline between Marseille andGibraltar or along the Thyrrennean coast.SO2 (Fig. 7c) is mainly emitted via indus-



0 4 8 12 16 20 240

4

8

12

16

20

24

Time (UTC)

Win

d sp

eed

(m s

−1 ) (a)

0 4 8 12 16 20 24−90

−60

−30

0

30

60

90

Time (UTC)

Win

d di

rect

ion

()

(b)

0 4 8 12 16 20 240

5

10

15

20

Time (UTC)

Tem

pera

ture

(C

) (c)

0 4 8 12 16 20 240

25

50

75

100

Time (UTC)

Rel

ativ

e hu

mid

ity (

%)

(d)

0 4 8 12 16 20 240

0.5

1

1.5

2

Time (UTC)

Win

d st

ress

(N

m−2 ) (e)

Fig. 6. Surface wind speed (a) and direction (b), temperature (c), relative humidity (d) and surfacestress (e) as a function of time from the ASIS buoy measurements (dashed line) (see diamond inFig. 1c) and from the MM5 simulations (solid line).

trial activities and domestic combustion, asshown by the more scattered emission pat-tern. The emissions associated with mar-itime traffic are significant. CO and coarseprimary particulate matter (PPM, of sizeless than 10 µm and more than 2.5 µm)emissions are shared by traffic and varioussources of combustion. These sources aresignificant in the main western Mediter-ranean cities as shown in Fig. 7 (b and d).

4.2 Analysis at the scale of the Gulf ofLion

On the shore of the Gulf of Lion, thearea around the city of Marseille and theBerre pond, in southeastern France is asource region of large urban and industrialemission with high occurrence of photo-chemical pollution events (Drobinski et al.2006). The aerosol optical depth (AOD)at 865 nm derived from the SeaWIFS passover the Gulf of Lion at 1132 UTC on 24March 1998 (obtained from the SeaDASsoftware) is shown in Fig. 8 (c). Fig-ure 8 evidences that most of the west-



0 5 10

38

40

42

44

46

48(a)

Lat

itude

(°N

)

0 5 10

38

40

42

44

46

48(b)

0 5 10

38

40

42

44

46

48(c)

Longitude (°E)

Lat

itude

(°N

)

0 5 10

38

40

42

44

46

48(d)

Longitude (°E)

0.5

1

1.5

2

2.5

x 1012

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2x 10

11

0.5

1

1.5

2

2.5

3

3.5

x 1010

0.5

1

1.5

2

2.5

3

3.5

4

4.5

x 1011

Fig. 7. EMEP inventory emissions in Mg/(50 km × 50 km)/year, adjusted on the grid of the domain 3at 1200 UTC. Panels a, b, c and d represent the concentrations of NO (a), PPM coarse particles (≥2.5 µm and ≤ 10 µm) (b), SO2 (c) and CO (d) during a labor day of March 1998.

ern Mediterranean was covered by clouds,with the notable exception of Gulf of Lion,which was partly cloud-free, on account ofthe strong Mistral jet blowing in the region.AODs at 865 nm could only be retrievedin the cloud-free part of the Gulf of Lion,but displays a minimum south of the city ofMontpellier (0.11 around 4◦E) and a maxi-mum south of the Fos/Berre industrial area(0.18 around 4.7◦E) along leg AF of theARAT aircraft (Fig. 1). In the Mistral re-gion (leg FC), the AOD at 865 nm variedfrom over 0.2 near the coast (Fos/Berre re-gion and further east) to 0.17 around way-point F (Fig. 1). The aerosol distributionsimulated by CHIMERE using the recentlydeveloped methodology by Hodzic et al.(2004), can be compared to the aerosol op-tical depth retrieved from SeaWIFS in thecloud-free regions. Figure 8b shows theAOD at 865 nm simulated with CHIMEREover the domain documented by SeaWIFS.

The location of the Marseille/Fos-Berreaerosol plume is reproduced by CHIMERE(shifted by about 20 km to the east) with anAOD of about 0.15, as well as the aerosolplume from the industrial city of Toulon.The plume emerging from the Aude valleychannelling the Tramontane as well as theplume skirting the eastern Spanish coast(Fig. 8a) are not visible in the SeaWIFSdata because these regions are covered byclouds preventing from reliable retrieval ofaerosol products (Fig. 8b, c). Despite theglobal good agreement, the AOD simu-lated in some plumes can be half smallerthan the observed AOD.

In the afternoon March 24, 1998, thevertical structure of aerosols as well asthe ABL were documented by the lidarLEANDRE-2 along legs AF, FC and CE(Fig. 9a) as detailled by Flamant (2003).Lidar measurements on leg AF (Fig. 9b)show an internal boundary layer develop-



AO

D (

865

nm)

0

0.1

0.2

0.3

0.4

2 4 6 8 1039

41

43

45

Longitude (°E)La

titud

e (°

N)

(a)

3 4 5 6 741

42

43

Longitude (°E)

Latit

ude

(°N

)

(b)

AO

D (

865

nm)

0

0.1

0.2

0.3

0.4

3 4 5 6 741

42

43

Longitude (°E)

Latit

ude

(°N

)

(c)

Fig. 8. Panel a: simulated aerosol optical depth (AOD) at 865 nm derived from CHIMERE simulation(domain 2). The rectangle indicates the zoomed area shown in panels (b) and (c); Panel b: simulatedAOD at 865 nm over the Gulf of Lions multiplied by a factor of 2 with the cloud mask derivedfrom the Sea-viewing Wide Field-of-view Sensor (SeaWIFS) observations; Panel c: AOD at 865 nmderived from SeaWIFS measurements.

ing from the coast (within the advectedcontinental ABL) and reaching a depth of1200 m. Close to way-point F, at approx-imately 42.6◦N, the marine ABL struc-ture characteristics over the sea changed: it is observed to be shallower (700 min Fig. 9b) and is characterized by largervalues of atmospheric reflectivity. This isconfirmed by measurements made on legFC (Fig. 9c), along which the marine ABLis characterized by values of atmosphericreflectivity similar to those observed nearway-point F. Also, the marine ABL is ob-served to remain shallow, its depth grad-ually decreasing from 700 to 500 m withthe distance to the coast. As discussed inFlamant (2003), the larger atmospheric re-flectivity values in the eastern part of thesampled region is thought to be related tolarger concentrations of pollution aerosolfrom the city of Marseille and the indus-trial petrochemical complex of Fos/Berre.On leg CE (Fig. 9d), close to the coast,the ABL structure is similar to that ob-

served along FC (i.e., shallow). The depthof the ABL increases westward from 500to 1200 m (between 43.25 and 43 ◦N) withreflectivity values on the order of those ob-served on leg CE. Further to the west, acontinental ABL is observed in the shel-tered region in the lee of the Massif Cen-tral, reflectivity values being significantlysmaller than to the east of 43 ◦N. In thisregion, lidar measurements evidence thepresence of gravity waves inside the con-tinental ABL and above, which are com-mon features during Mistral events (Cac-cia et al. 2004; Drobinski et al. 2005).The vertical structure of the aerosol dis-tribution in the atmosphere derived fromLEANDRE-2 reflectivity is used to vali-date the CHIMERE simulations along theARAT legs. To do that, we use the method-ology by Hodzic et al. (2004) that involvessimulating the lidar backscattering profilesfrom the model relative humidity and con-centration outputs (optical properties vary-ing with chemical composition and mass



vertical distribution). Figures 9 and 10show the simulated lidar reflectivity andrelative humidity along legs AF, CE andCF. To obtain a quantitative comparisonbetween the measured and simulated re-flectivities, the simulated lidar profiles arenormalized by the measured values in theaerosol-free layer (at about 3.8 km in thetroposphere). Figure 10 shows a fairly hor-izontally homogeneous relative humidityfield meaning that the lidar reflectivity canbe directly related to the aerosol concentra-tion without any modulation of the aerosolsize by hygroscopic effect. Despite somedifferences in the lidar reflectivity field,the simulation displays many similar fea-tures. The main difference is the thin free-aerosol layer (reflectivity values of about50 a.u.) just above the ABL (reflectivityvalues larger than 150 a.u.) which is notreproduced by CHIMERE. Otherwise, thesecond layer above the ABL (the advectedcontinental ABL with reflectivity values ofabout 100-120 a.u.) is accurately simulatedfor all legs. The transition between themarine ABL and the advected continentalABL is thus too smooth in the simulationprobably because of too large diffusivitywhich prevents the model to predict thisaerosol-free layer between the marine ABLand the advected continental ABL. Alongleg CF, the aerosol backscatter is homoge-neous in the ABL both in the observationsand the simulation since the aircraft fliesalong the aerosol plume (Fig. 8), with adeepening of the ABL from east to west(the CHIMERE simulation, however, over-predicts by about 500 m the ABL depthwest of leg CF). Along legs CE and AF,the aerosol backscatter in the ABL variesalong the legs, with larger reflectivity onthe eastern part of the leg. On leg CE,the CHIMERE model produces large lidarreflectivity east of 3.9◦E. The location ofthe aerosol plume is accurately reproducedwith however a more diffuse plume edgebetween 3.8 and 4.3◦E in the LEANDRE-2 observations and smaller simulated re-flectivity east of 4.3◦E. West of 3.8◦E, theaerosol content decreases. Along leg AF,the model accurately reproduces the heav-ily loaded aerosol plumes east of 4.2◦Eand west of 3.8◦E. Between 3.8 and 4.2◦E,CHIMERE model simulates the incursion

of a low aerosol streak within the ABL, butis not able to reproduce its fine-scale struc-ture and the simulated reflectivity is largerthan the measured reflectivity.

To assess the origin air masses responsi-ble for the aerosol variability in ABL in theGulf of Lion area, we computed back tra-jectories ending at various locations in theaerosol plume documented by SeaWIFS(Fig. 11) and LEANDRE-2 (Figs. 12 and13), using a lagrangian particle dispersionmodel (Menut et al. 2000). The trajectoriesuse the MM5 wind fields. During the con-vective period, air parcels are mixed withinthe ABL. Hence neither constant-altitudetrajectories nor trajectories following themean three-dimensional wind would be re-alistic. In order to overcome this problem,air trajectories are assumed to undergo arandom altitude change every hour duringthe convective period. They were nev-ertheless bounded to stay between 0 and1000 m (tests with the upper boundary of2000 and 3000 m have been conductedwithout any impact). Fifty back trajecto-ries are calculated in this way (due to theconvective processes, an ending point doesnot correspond to the same origin, In or-der to take these various origins into ac-count, 50 trajectories are studied). Fig-ure 11 (a and b) shows that the air massesfollowing similar trajectories and endingon the western edge of the aerosol plume at1100 UTC can be clustered in two groups:one cluster follows the Aude valley chan-nelling the Tramontane flow, the other fol-lows the western boundary of the Mistralflow. There is no major urban and indus-trial aerosol sources along the two trajec-tories, except maybe the cities of Toulouseand Montpellier (Fig. 1b and 7). Figure 11(c and d) shows that the center and east-ern regions of the aerosol plume originatefrom the Marseille/Fos-Berre area but alsofrom further sources: To the north, emis-sions from Lyon are transported long rangeby the Mistral (see also Corsmeier et al.2005 for a summer Mistral) and to theeast, a cross-Alpine flow carries the emis-sions from the heavily industrialized re-gions of the Po Valley in Italy. The regionsof Marseille/Fos-Berre, Lyon and Torinoare major emission sources compared toToulouse or Montpellier, thus explaining



3.8 4 4.2 4.4 4.6

1

2

3

Longitude (°E)

Hei

ght (

km)

(a) Leg CE

CHIMERE simulation

3.8 4 4.2 4.4 4.6

1

2

3

Longitude (°E)

Hei

ght (

km)

(b) Leg CE

LEANDRE−2 measurements

42.4 42.6 42.8 43 43.2

1

2

3 (c) Leg CF

Latitude (°E)

Hei

ght (

km)

42.4 42.6 42.8 43 43.2

1

2

3 (d) Leg CF

Latitude (°E)

Hei

ght (

km)

3.8 4 4.2 4.4 4.6

1

2

3 (e) Leg AF

Longitude (°E)

Hei

ght (

km)

Reflectivity (a.u.)50 100 150

3.8 4 4.2 4.4 4.6

1

2

3 (f) Leg AF

Longitude (°E)

Hei

ght (

km)

Fig. 9. Atmospheric reflectivity at 0.73 nm as obtained from LEANDRE-2 (right column) andCHIMERE (left column) along leg CE (a and b), leg CF (c and d), and leg AF (e and f) on theafternoon of March 24 1998. Way points A, C, and E are shown in Fig. 1c. The aerosol-laddenmarine ABL is characterized by high reflectivity values, while the advected continental ABL exhibitsreflectivity values comprised between 80 and 1200 a.u. The free troposphere above the ABL, ischaracterized by a reflectivity values of 80 a.u. or less.

why the western region of the plume is op-tically thinner than the middle and easternregions of the plume.

The back plumes ending along theARAT legs at two different altitudes (50 m

within the ABL, 2000 m within the ad-vected aerosol-rich continental ABL abovethe marine ABL) also explain the differ-ences in aerosol concentration observedby LEANDRE-2. At 50 m (and higher



Hei

ght (

km)

Longitude (oE)

(a)

4 4.2 4.4 4.60

1

2

3

4

Hei

ght (

km)

Latitude (oN)

(b)

42.6 42.8 43 43.20

1

2

3

4

Hei

ght (

km)

Longitude (oE)

(c)

3.8 4 4.2 4.4 4.60

1

2

3

4

Rel

ativ

e H

umid

ity (

%)

10

20

30

40

50

60

Fig. 10. Vertical profiles of relative humidity simulated with MM5 and extracted from domain 2along legs AF (a), CE (b) and CF (c).

within the ABL, not shown), large-scaletransport from Lyon is visible on thefour clusters of back trajectories. How-ever, at ending points A and E, the airmasses do not flow over the Marseille/Fos-Berre area or other major cities, explain-ing the lower aerosol content observed byLEANDRE-2 within the ABL. On the con-trary, the air masses ending at points Cand F are enriched in aerosols as theypass over the Marseille/Fos-Berre area ex-plaining the higher reflectivity values inthe ABL. Along leg CF, the ARAT fol-lows the Marseille/Fos-Berre plume ex-plaining the homogeneous concentrationof aerosols along the leg.

At 2000 m, the aerosols of the layer justabove the marine ABL (reflectivity valuesof about 100 a.u.) originate from remotesources (e.g. Lyon) which are then trans-

ported over a long range (Corsemeier et al.2005) (Fig 13). One exception is at endingpoint C which is just above this aerosol-rich layer, in the aerosol-free layer (Fig. 9).Figure 13c shows that the air mass flowsover the advected continental ABL abovethe western Alps and originate from thenorthern Alps. The absence of significantaerosol sources in this area explains thevery low reflectivity values at 2000 m atending point C. One can finally notice theabsence of trans-alpine transport between1600 and 1700 UTC contrary to the situa-tion at 1100 UTC. This is the result of anevolution of the regime of the flow imping-ing the Alpine ridge: at 1100 UTC, part ofthe upstream air mass flows over the moun-tain whereas at 1600 and 1700 UTC, theair mass is mostly blocked by the Alps andflows around. In Fig. 2, it is illustrated by



0 2 4 6 8 10 1240

42

44

46

48

Lat

itude

(o N

)

(a)

0 2 4 6 8 10 1240

42

44

46

48(b)

0 2 4 6 8 10 1240

42

44

46

48

Longitude (oE)

Lat

itude

(o N

)

(c)

0 2 4 6 8 10 1240

42

44

46

48

Longitude (oE)

(d)

Fig. 11. Back trajectories ending at 3.3◦E/42.5◦N (a), 3.39◦E/43◦N (b), 3.95◦E/43.16◦N (c) and4.75◦E/42.6◦N (d) (shown in Fig. 8). The altitude of the ending points is 50 m. The ending timeof the back trajectories is 1100 UTC which is the time of the passage of SeaWIFS over the westernMediterranean. The wind fields used to compute these back trajectories are from the 9-km meshsize domain. Filled black dots indicate the cities of Barcelona, Toulouse, Montpellier, Marseille/Fos-Berre, Lyon, Torino and Milano.

the larger extension of the western Alpinewake in the afternoon.

4.3 Aerosol distribution over the west-ern Mediterranean from the MM5-CHIMERE model

Insight into the aerosol distribution, ori-gin and composition over the westernMediterranean can be obtained from theCHIMERE aerosol products, in close re-lationship with the mesoscale dynamicalprocesses accessible from the validatedMM5 meteorological fields. In the fol-lowing, the aerosol distribution, origin andcomposition will be discussed as a functionof the position and intensity of the ”eye” ofthe Genoa cyclone on 24 March.

During the morning period, the Genoacyclone is located immediately to the

east of Corsica, the surface pressurelow gradually decreasing from 1014 to1010 hPa. Over the southwestern regionof the Mediterranean basin, large aerosolloading can be attributed to ship emissions(see Fig. 7 and 14a and b). South of Sar-dinia, the air circulating around the Genoasurface low brings ship emissions to south-ern Italy. North of this band, most ofthe aerosol loading over the MediterraneanSea is of local origin (inter-continental shiptracks and coastal urban and industrializedarea such as Barcelona, Marseille/Fos-Berre, Toulon, see Fig. 7 and Fig. 14a andb), with one exception along the Aude val-ley where the Tramontane transports off-shore continental aerosols. The plume overToulon stagnates in the western Alps waketrailing downstream across the Alps, anarea of very weak wind. Finally, the close



0 2 4 6 8 10 1240

42

44

46

48

Lat

itude

(o N

)(a)

0 2 4 6 8 10 1240

42

44

46

48(b)

0 2 4 6 8 10 1240

42

44

46

48

Longitude (oE)

Lat

itude

(o N

)

(c)

0 2 4 6 8 10 1240

42

44

46

48

Longitude (oE)

(d)

Fig. 12. Back trajectories ending at 3.9◦E/43.4◦N (way-point A) (a), 4.65◦E/42.5◦N (way-point F)(b), 4.7◦E/43.2◦N (way-point C) (c) and 3.88◦E/42.75◦N (way-point E) (d). The back trajectoriesending points correspond to the starting and ending points of the ARAT legs and the aerosol plumesdetected by the lidar LEANDRE-2 on board the ARAT. The altitude of the ending points is 50 m. Theending times of the back trajectories are 1600 and 1700 UTC for the first and second row, respectively.Filled black dots indicate the cities of Barcelona, Toulouse, Montpellier, Marseille/Fos-Berre, Lyon,Torino and Milano.

location of the Genoa cyclone from Cor-sica and the absence of strong winds dur-ing the morning period lead to the fact thatthe region north and northwest of Corsicais only influenced by marine clean air.

During the afternoon period, theGenoa cyclone intensifies (from 1010to 1004 hPa) and moves southeastward,about 300 km east of Sardinia over theTyrrhenian Sea. The Tramontane hasnearly vanished so that the polluted air ofthe morning stagnates in the Aude valley.Conversely, the Mistral has intensified andmost of the aerosols exported offshoreoriginates from central France and theMarseille/Fos-Berre plumes. Moreover,the Ligurian outflow brings in the aerosolsemitted from the major industrializedItalian cities (Milano and Torino) over

the western part of the Mediterraneanbasin. This is consistent with aerosoloptical depth values of 0.1 and 0.3 at 870and 440 nm, respectively, measured atIspra (Italy) with the photometer of theAERONET network and with spectraldistribution revealing high accumula-tion mode aerosol concentration. TheCHIMERE AOD at 865 nm at Ispra isabout 0.09. A narrow band of clean airseparates the two major plumes extendingmore than 300 km over the sea. Thiscleaner air corresponds to the westernAlps wake and, associated with large hori-zontal shear of wind, acts as a dynamicalbarrier, preventing any lateral exchangesbetween the Mistral and the Ligurianoutflow. The southeastward displacementand intensification of the Genoa cyclone



0 2 4 6 8 10 1240

42

44

46

48

Lat

itude

(o N

)

(a)

0 2 4 6 8 10 1240

42

44

46

48(b)

0 2 4 6 8 10 1240

42

44

46

48

Longitude (oE)

Lat

itude

(o N

)

(c)

0 2 4 6 8 10 1240

42

44

46

48

Longitude (oE)

(d)

Fig. 13. Same as Fig. 12 for back trajectories ending at 2000 m MSL.

increases the meridian extension of theMistral that sweeps away the band ofaerosols emitted from the ships that skirtthe North African coast in the morning(Fig. 14c and d).

Despite the difficulty in validating thesimulated composition of the aerosol load-ing over the basin, measurements of ni-trates available in Netherlands and Italyclearly show the development of an ammo-nium nitrate episode over the western Eu-rope from March 22 to 31 (see measuredand simulated values in Table 1, NL09,IT04). These episodes are often observedin the late winter during such meteoro-logical conditions with high pressure overGreat Britain and central Europe (Bessag-net et al., 2005). The large nitrogen ox-ides emissions in the Netherlands, Bel-gium, Germany and northern France pro-duce by oxidation a large amount of ni-tric acid which is then neutralized by am-monia emitted in the same regions andform ammonium nitrate in the particulatephase. Moreover, the coarse simulation

displays the formation of a sulfate bub-ble in Great Britain on March 22 goingto southern France on March 23 and 24(not shown). In the same time, the sul-fate concentration increases when crossingwestern France (Normandy) due to highsulfur oxide emissions. This behavior issupported by sulfate observations in threeEMEP background sites in Great Britainand France (see Table 1, GB02, GB14 andFR03). During the development of thisepisode, in particular on March 24, accord-ing to the model results, the Genoa cyclonedrives the pollution plume towards south-ern and western France. Afterwards, theammonium nitrate is advected through theAude and Rhone valleys to the Mediter-ranean basin. A similar ammonium nitrateplume is also transported from the Po val-ley (Fig. 14). The formation of ammoniumnitrate is favored by low temperatures insouthern France which reach a maximumof about 10◦C near the surface on March24 at 1200 UTC and avoid the evaporationof ammonium nitrate.



Fig. 14. Aerosol loading with respect to the layer hight, over the western Mediterranean basin. Pan-els a, b, c and d display the particulate concentrations fields modelled with CHIMERE on domain 2at 0600, 0900, 1200, 1500, 1800 and 2100 UTC, respectively. Solid lines represent the coastlines andelevations beyond 500 m.

5 Conclusions

Finally, this article presents a first at-tempt to characterize the aerosol distribu-

tion over the western Mediterranean basinduring a Mistral event from the combi-nation of innovative instrumentation (pas-



Table 1. Nitrate concentrations measured/simulated in the Netherlands (NL09) and Italy (IT03)and sulfate concentrations measured/simulated in Great Britain (GB02 and GB14) and France(FR03) between March 22 and 31. The stations are from the EMEP measurement network and theconcentrations are daily averaged. The simulated values are given between brackets.

Nitrates (µg m−3) Date NL09 IT04(6.28◦E; 53.33◦N) (8.62◦E; 45.82◦N)

March 22, 1998 1.28 (0.79) 4.43 (7.11)March 23, 1998 3.23 (1.96) 5.00 (9.35)March 24, 1998 5.49 (4.41) 4.21 (3.12)March 25, 1998 7.71 (13.66) 9.83 (6.24)March 26, 1998 7.26 (8.22) 17.80 (22.66)March 27, 1998 8.99 24.98March 28, 1998 8.99 27.90March 29, 1998 11.69 21.17March 30, 1998 9.48 23.25March 31, 1998 10.23 18.11

Sulfates (µg m−3) Date GB02 GB14 FR03(-3.20◦E; 55.32◦N) (-0.80◦E; 54.33◦N) (1.38◦E; 46.13◦N)

March 22, 1998 8.34 (3.62) 4.95 (3.16) 0.30 (1.59)March 23, 1998 8.97 (8.58) 5.82 (4.61) 5.61 (2.36)March 24, 1998 1.05 (1.33) 6.30 (3.57) 13.20 (3.18)?

March 25, 1998 1.14 (0.80) 2.76 (2.02) 1.83 (2.55)March 26, 1998 0.81 (0.63) 1.62 (1.04) 3.00 (2.01)March 27, 1998 2.61 3.30 1.83March 28, 1998 5.85 6.60 4.71March 29, 1998 5.52 6.84 2.40March 30, 1998 1.20 3.00 5.07March 31, 1998 0.48 1.14 3.69

?: the simulated sulfate plume with concentrations of about 8-9 µg m−3 is about 50 km to the westof the measurement station, thus explaining the large difference.

sive and active remote sensors) and theaerosol products of the fine-scale chem-istry transport model CHIMERE forcedwith the non-hydrostatic mesoscale modelMM5. This issue is motivated by (1)the potential impact of aerosols on theMediterranean radiate budget, the watercycle (through the modulation of precip-itations) and the marine ecosystems; and(2) the very frequent occurrence of Mistral-type weather regimes (5 to 15 days permonth all year long). After validating thedynamical and chemical simulations, thispaper shows that:

1. the Mistral contributes to export faroffshore a large amount of aerosolsemitted over continental Europe (inparticular ammonium nitrate in theparticulate phase and sulfates) andalong the shore from the indus-trialized and urban areas of Fos-Berre/Marseille, as well as from the

Po valley. The Genoa surface lowcontributes to advect the aerosolsalong a cyclonic trajectory whichskirts the North African coast andreach Italy.

2. due to the time evolution of the Genoacyclone position, the aerosol con-centration pattern is very unsteady,as the Tramontane prevails in themorning hours of March 24, leavingplace to the Mistral and an unusuallystrong Ligurian outflow in the after-noon. The wakes trailing downstreamthe Massif Central and the Alps pre-vent any horizontal diffusion of theaerosols and can, in some occasions,contribute to aerosol stagnation.

Future work will be dedicated to assessthe representativity of the aerosol distri-bution during the very numerous Mistralevents.



Acknowledgements. This work was conductedat Service d’A eronomie and Laboratoire deM et eorologie Dynamique of the Institut PierreSimon Laplace. The authors would like tothank T. Elias, G. Bergametti, S. Bastin andC. Moulin for fruitful discussion; J.L. Mongefor technical support; M.C. Lanceau for helpin collecting the referenced papers; and D.Hauser for the coordination of the experiment.This research has received funding from theAgence de l’Environnement et de la Ma ıtrise del’Energie (ADEME) and support from M et eo-France that provided the data from the opera-tional radiosoundings and meteorological sur-face stations.

References

Aebischer, U., and Schar, C.: Low-level poten-tial vorticity and cyclogenesis to the lee ofthe Alps, J. Atmos. Sci., 55, 186–207, 1998.

Alpert, P., Tsidulko, M., Krichak, S., and Stein,U.: A multi-stage evolution of an ALPEXcyclone, Tellus, 48, 209–220, 1996.

Bastin, S., Drobinski, P., Gu enard, V., Caccia,J.L., Campistron, B., Dabas, A.M., Delville,P., Reitebuch, O., and Werner, C.: On theinteraction between sea-breeze, and summerMistral at the exit of the Rh one valley, Mon.Wea. Rev., 134, 1647–1668, 2006.

Bastin, S., and Drobinski, P.: Sea breeze in-duced mass transport over complex terrain insoutheastern France: a case study, Q. J. R.Meteorol. Soc., 132, 405–423, 2006.

Benitez-Nelson, C.: The biogeochemical cy-cling of phosphorus in marine systems,Earth-Science Review, 51, 109–135, 2000.

Bergametti, G., Remoudaki, E., Losno, R.,Steiner, E., Chatenet, B., and Buat-Menard,P.: Source, transport and deposition of atmo-spheric phosphorus over the NorthwesternMediterranean, J. Atmos. Chem., 14, 501–513, 1992.

Bessagnet, B., Hodzic, A., Blanchard, O., Lat-tuati, M., Le Bihan, O., Marfaing H., andRouil, L.: Origin of particulate matter pol-lution episodes in wintertime over the ParisBasin, Atmos. Env., 39, 6159–6174, 2005.

Bessagnet, B., Hodzic, A., Vautard, R., Beek-mann, M., Cheinet, S., Honore, C., Li-ousse, C., and Rouil, L.: Aerosol modelingwith CHIMERE-First evaluation at continen-tal scale, Atmos. Env., 38, 2803–2817, 2004.

B ethoux, J.P.: Oxygen consumption, new pro-duction, vertical advection and environmen-tal evolution in the Mediterranean Sea, Deep-Sea Research, 36, 769–781, 1989.

B ethoux, J.P., Morin, P., and Ruiz-Pino, D.:Temporal trends in nutrient ratios: chem-ical evidence of Mediterranean ecosystemchanges driven by human activity, Deep-SeaResearch II, 49, 2007–2016, 2002.

Buzzi, A., and Tibaldi, A.: Cyclogenesis in thelee of the Alps: A case study, Q. J. R. Mete-orol. Soc., 104, 271–287, 1978.

Caccia J.L., Gu enard, V., B enech, B., Camp-istron, B., and Drobinski, P.: Vertical ve-locity and turbulence aspects during Mistralevents as observed by UHF wind profilers,Ann. Geophys., 22, 3927–3936, 2004.

Chiriaco, M., Vautard, R., Chepfer, H., Haeffe-lin, M., Wanherdrick, Y., Morille, Y., Protat,A., and Dudhia, J.: The ability of MM5 tosimulate ice clouds: systematic comparisonbetween simulated and measured fluxes andlidar/radar profiles at SIRTA atmospheric ob-servatory, Mon. Wea. Rev., 134, 897–918,2006.

Corsmeier, U., Behrendt, R., Drobinski, P., andKottmeier, C.: The Mistral and its effect onair pollution transport and vertical mixing,Atmos. Res., 74, 275–302, 2005.

Drennan, W., Graber, H.C., Hauser, D., andQuentin, C.: On the wave age dependence ofwind stress over pure wind seas, J. Geophys.Res., 108, doi:10.1029/2000JC000715,2003.

Drobinski, P., Saıd, F., Ancellet, G., Arteta,J., Augustin, P., Bastin, S., Brut, A., Cac-cia, J.L., Campistron, B., Cautenet, S., Co-lette, A., Coll, I., Cros, B., Corsmeier, U.,Dabas, A., Delbarre, H., Dufour, A., Du-rand, P., Gu enard, V., Hasel, M., Kalthoff,N., Kottmeier, C., Lemonsu, A., Lasri, F. Lo-hou, F., Masson, V., Menut, L., Moppert, C.,Peuch, V.H., Puygrenier, V., Reitebuch, O.,and Vautard, R.: Regional transport and dilu-tion during high pollution episodes in south-ern France: Summary of findings from theESCOMPTE experiment, J. Geophys. Res.,in press, 2006.

Drobinski, P., Bastin, S., Gu enard, V., Caccia,J.L., Dabas, A.M., Delville, P., Protat, A.,Reitebuch, O., and Werner, C.: Summer Mis-tral at the exit of the Rh one valley, Q. J. R.Meteorol. Soc., 131, 353–375, 2005.

Drobinski, P., Flamant, C., Dusek, J., Flamant,P.H., and Pelon, J.: Observational evidenceand modeling of an internal hydraulic jumpat the atmospheric boundary-layer top duringa Tramontane event, Bound.-Layer Meteor.,98, 497–515, 2001a.

Drobinski, P., Dusek, J., and Flamant, C.: Di-agnostics of hydraulic jump and gap flowin stratified flows over topography, Bound.-



Layer Meteor., 98, 475–495, 2001b.Dudhia, J.A.: Nonhydrostatic version of the

Penn State-NCAR mesoscale model: valida-tion tests and simulation of an Atlantic cy-clone and cold front, Mon. Wea. Rev., 121,1493–1513, 1993.

Egger, J.: Numerical experiments on the cyclo-genesis in the Gulf of Genoa, Contrib. At-mos. Phys., 45, 320–346, 1972.

Eymard, L., Weill, A., Bourras, D., Guerin,C., Le Borgne, P., and Lefevre, J.M.: Useof ship mean data for validating model andsatellite flux fields during the FETCH exper-iment, J. Geophys. Res., 108, doi:10.1029/2001JC001207, 2003.

Flamant, C.: Alpine lee cyclogenesis influenceon air-sea heat exchanges and marine atmo-spheric boundary layer thermodynamics overthe western Mediterranean during a Tramon-tane/Mistral event, J. Geophys. Res., 108,doi:10.1029/2001JC001040, 2003.

Flamant, C., Pelon, J., Hauser, D., Quentin,C., Drennan, W.M., Gohin, F., Chapron, B.,and Gourrion, J.: Analysis of surface windand roughness length evolution with fetchusing a combination of airborne lidar andradar measurements, J. Geophys. Res., 108,doi:10.1029/2002JC001405, 2003.

Flamant, C., and Pelon, J.: Atmosphericboundary-layer structure over the Mediter-ranean during a Tramontane event, Q. J. R.Meteorol. Soc., 122, 1741–1778, 1996.

Georgelin, M., and Richard, E.: Numeri-cal simulation of flow diversion around thePyren ees: a Tramontana case study, Mon.Wea. Rev., 124, 687–700, 1996.

Grell, G.A.: Prognostic evaluation of assump-tions used by cumulus parameterizations,Mon. Wea. Rev., 121, 764–787, 1993.

Grell, G.A., Dudhia, J., and Stauffer, D.R.:A description of the fifth-generationPenn State/NCAR mesoscale model(MM5), NCAR Technical Note, NCAR/TN-398+STR, 122 pp, 1995.

Grotjahn, R., and Wang, C.H.: On the sourceof air modified by surface fluxes to enhancefrontal cyclone development, Ocean-Air In-teract., 1, 257–288, 1989.

Gu enard, V., Drobinski, P., Caccia, J.L.,T edeschi, G., and Currier, P.: Dynamics ofthe MAP-IOP 15 severe Mistral event: obser-vations and high-resolution numerical simu-lations, Q. J. R. Meteorol. Soc., 132, 757–778, 2006.

Gu enard, V., Drobinski, P., Caccia, J.L., Camp-istron, B., and B enech, B.: An observationalstudy of the mesoscale Mistral dynamics,Bound.-Layer Meteor., 115, 263–288, 2005.

Guerzoni, S., Chester, R., Dulac, F., Herut,B., Loye-Pilot, M.D., Measures, C., Mignon,C., Molinaroli, E., Moulin, C., Rossini, P.,Saydam, C., Soudine, A., and Ziveri, P.:The role of atmospheric deposition in thebiogeochemistry of the Mediterranean Sea,Progress in Oceanography, 44, 147–190,1999.

Guieu, C., Bozec, Y., Blain, S., Ridame,C., Sarthou, G., and Leblond, N.: Im-pact of high Saharan dust inputs ondissolved iron concentrations in the Mediter-ranean Sea, Geophys. Res. Let., 29,doi:10.1029/2001GL014454, 2002.

Guo, Y.R., and Chen, S.: Terrain and land usefor the fifth-generation Penn State/NCARmesoscale modeling system (MM5), NCARTechnical Note, NCAR/TN-397+IA, 114 pp,1994.

Hauser, D., Branger, H., Bouffies-Cloch e, S.,Despiau, S., Drennan, W.M., Dupuis, H.,Durand, P., Durrieu de Madron, X., Estour-nel, C., Eymard, L., Flamant, C., Graber,H.C., Gu erin, C., Kahma, K., Lachaud, G.,Lefevre, M.P., Pelon, J., Pettersson, H.,Piguet, B., Queffeulou, P., Tailliez, D., Tour-nadre J., and Weill, A.: The FETCH exper-iment: an overview, J. Geophys. Res., 108,doi:10.1029/2001JC001202, 2003.

Hong, S.Y., and Pan, H.L.: Nonlocal boundarylayer vertical diffusion in a medium-rangeforecast model, Mon. Wea. Rev., 124, 2322–2339, 1996.

Hodzic, A., Chepfer, H., Vautard, R., Chazette,P., Beekmann, M., Bessagnet, B., Chatenet,B., Cuesta, J., Drobinski, P., Goloub, P., Ha-effelin, M., and Morille, Y.: Comparisonof aerosol chemistry-transport model simu-lations with lidar and sun-photometer obser-vations at a site near Paris, J. Geophys. Res.,109, D23201, doi:10.1029/2004JD004735,2004.

Jans a, A.: Distribution of the mistral: A satel-lite observation, Meteorol. Atmos. Phys., 36,201–214, 1987.

Lelieveld, J., Berresheim, H., Borrmann, S.,Crutzen, P.J., Dentener, F.J., Fischer, H., Fe-ichter, J., Flatau, P.J., Heland, J., Holzinger,R., Korrmann, R., Lawrence, M.G., Levin,Z., Markowicz, K.M., Mihalopoulos, N.,Minikin, A., Ramanathan, V., de Reus, M.,Roelofs, G.J., Scheeren, H.A., Sciare, J.,Schlager, H., Schultz, M., Siegmund, P.,Steil, B., Stephanou, E.G., Stier, P., Traub,M., Warneke, C., Williams, J., and Ziereis,H.: Global air pollution crossroads overthe Mediterranean, Science, 298, 794–799,2002.



Marmer, E., and Langmann, B.: Impact ofship emissions on the Mediterranean sum-mertime pollution and climate: A regionalmodel study, Atmos. Env., 39, 4659–4669,2005.

Mayencon, R.: M et eorologie pratique, EditionsMaritimes D’outre-mer, 1982

Menut, L., Vautard, R., Flamant, C., Abon-nel, C., Beekmann, M., Chazette, P., Fla-mant, P.H., Gombert, D., Gu edalia, D., Kley,D., Lefebvre, M.P., Lossec, B., Martin,D., M egie, G., Perros, P., Sicard, M., andToupance, G.: Measurements and modellingof atmospheric pollution over the Paris area:an overview of the ESQUIF project, Ann.Geophys., 18, 1467–1481, 2000.

Migon, C., Copin-Mont egut, G., El egant,L., and Morelli, J.: Etude de l’apportatmosph erique en sels nutritifs au mi-lieu c otier m editerran een et implicationsbiog eochimiques, Oceanologica Acta, 12,187–191, 1989.

Millot, C.: Wind induced upwellings in the Gulfof Lions, Oceanol. Acta, 2, 261–274, 1979.

Mlawer, E.J., Taubman, S.J., Brown, P.D.,Iacono, M.J., and Clough, S.A.: Radia-tive transfer for inhomogeneous atmosphere:RRTM, a validated correlated-k model forthe longwave, J. Geophys. Res., 102 (D14),16663–16682, 1997.

Moulin, C., Lambert, C.E., Dayan, U., Mas-son, V., Ramonet, M., Bousquet, P., Legrand,M., Balkanski, Y.J., Guelle, W., Marticorena,B., Bergametti, G., and Dulac, F.: Satel-lite climatology of African dust transportin Mediterranean atmosphere, J. Geophys.Res., 103, 137–144, 1998.

Moulin, C., Lambert, C.E., Dulac, F., andDayan, U.: Control of atmospheric export ofdust from North Africa by the North AtlanticOscillation, Nature, 387, 691–694, 1997.

Orieux, A., and Pouget, E.: Le Mistral: Con-tribution a l’ etude de ses aspects synoptiqueset r egionaux, Direction de la M et eorologie,Monographie 5, 1984

Pettr e, P.: On the problem of violent valleywind, J. Atmos. Sci., 39, 542–554, 1982.

Prospero, J.M., Barrett, K., Church, T., Den-tener, F., Duce, R.A., Galloway, J.N., LevyII, H., Moody, J., and Quinn, P.: Atmo-spheric deposition of nutrients to the NorthAtlantic Basin, Biogeochemistry, 35, 27–73,1996.

Reisner, J., Rasmussen, R.J., and Bruintjes,R.T.: Explicit forecasting of supercooled liq-uid water in winter storms using the MM5mesoscale model, Q. J. R. Meteorol. Soc.,124, 1071–1107, 1998.

Schar, C., and Smith, R.B.: Shallow-water flowpast isolated topography. Part I: Vorticityproduction and wake formation, J. Atmos.Sci., 50, 1373–1400, 1993.

Schmidt, H., Derognat, C., Vautard, R., andBeekmann, M.: A comparison of simulatedand observed ozone mixing ratios for thesummer of 1998 in Western Europe, Atmos.Envi., 35, 6277–6297, 2001.

Schneider, J., Borrmann, S., Wollny, A.G.,Blasner, M., Mihalopoulos, N., Oikonomou,K., Sciare, J., Teller, A., Levin, Z., andWorsnop, D.R.: Online mass spectromet-ric aerosol measurements during the MI-NOS campaign (Crete, August 2001), At-mos. Chem. Phys., 4, 65–80, 2004.

Sciare, J., Cachier, H., Oikonomou, K., Ausset,P., Sarda-Esteve, R., and Mihalopoulos, N.:Characterization of carbonaceous aerosolsduring the MINOS campaign in Crete, July-August 2001: a multi-analytical approach,Atmos. Chem. Phys., 3, 1743–1757, 2003.

Traub, M., Fischer, H., de Reus, M., Kormann,R., Heland, J., Ziereis, H., Schlager, H.,Holzinger, R., Williams, J., Warneke, C., deGouw, J., and Lelieveld, J.: Chemical char-acteristics assigned to trajectory clusters dur-ing the MINOS campaign, Atmos. Chem.Phys., 3, 459–468, 2003.

Vautard R., Martin, D., Beekmann, M.,Drobinski, P., Friedrich, R., Jaubertie,A., Kley, D., Lattuati, M., Moral, P.,Neininger, B., and Theloke, J.: Paris emis-sion inventory diagnostics from the ESQUIFairborne measurements and a chemistry-transport model, J. Geophys. Res., 108,8564, doi:10.1029/2002JD002797, 2003.

Vautard, R., Beekmann, M., Roux, J., andGombert, D.: Validation of a hybrid forecast-ing system for the ozone concentrations overthe Paris area, Atmos. Env., 35, 2449-2461,2001.

Vestreng, V.: EMEP/MSC-W Technical re-port. Review and Revision. Emission data re-ported to CLRTAP. MSC-W Status Report2003. EMEP/MSC-W Note 1/2003. ISSN0804-2446, 2003.

Wrathall, J.E.: The Mistral and forestfires in Provence-C ote d’Azur-South France,Weather, 40, 119–124, 1985.


tramontane/mistral eventmenut/pp/2006salameh-acpd-0318-ms.pdf · basin in order to provide insight...

Documents