The impact of dust storms on the Arabian Peninsula and the - ACPD

33 Pages·2014·6.48 MB·English

Atmos.Chem.Phys.Discuss.,14,19181–19245,2014 D
www.atmos-chem-phys-discuss.net/14/19181/2014/ isc
doi:10.5194/acpd-14-19181-2014 us
s
©Author(s)2014.CCAttribution3.0License. io
n
P
a
Thisdiscussionpaperis/hasbeenunderreviewforthejournalAtmosphericChemistry pe
r
andPhysics(ACP).PleaserefertothecorrespondingfinalpaperinACPifavailable.
|
The impact of dust storms on the Arabian D
is
c
u
Peninsula and the Red Sea ss
io
n
P
a
P.JishPrakash,G.Stenchikov,S.Kalenderski,S.Osipov,andH.Bangalath p
e
r
DivisionofPhysicalSciencesandEngineering,KingAbdullahUniversityofScienceand
|
Technology,Thuwal,SaudiArabia
D
Received:24February2014–Accepted:1July2014–Published:22July2014 isc
u
s
Correspondenceto:G.Stenchikov([email protected]) s
io
n
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.
P
a
p
e
r
|
D
is
c
u
s
s
io
n
P
a
p
e
r
19181 |
Abstract D
is
c
u
Locatedinthedustbelt,theArabianPeninsulaisamajorsourceofatmosphericdust. s
s
Frequentdustoutbreaksandsome15to20duststormsperyearhaveprofoundeffects io
n
on all aspects of human activity and natural processes in this region. To quantify the P
a
5 effect of severe dust events on radiation fluxes and regional climate characteristics, pe
r
wesimulatedthestormthatoccurredon18–20March2012usingaregionalweather
researchforecastmodelfullycoupledwiththechemistry/aerosolmodule(WRF-Chem). |
ThisstormsweptoveraremarkablylargeareaaffectingtheentireMiddleEast,North- D
is
Eastern Africa, Afghanistan and Pakistan. It was caused by a southward propagating c
u
10 cold front and associated winds activated the dust production in river valleys of the ssio
lower Tigris and Euphrates in Iraq, the coastal areas in Kuwait, Iran, and the United n
P
Arab Emirates, Rub al Khali, An Nafud and Ad Dahna deserts, and along the Red a
p
SeacoastonthewestsideoftheArabianPeninsula.Oursimulationresultscompare e
r
well with available ground-based and satellite observations. The total amount of dust
|
generatedbythestormreached93.76Mt.About80%ofthisamountdepositedwithin
15
D
thecalculationdomain.TheArabianSeaandPersianGulfreceived5.3Mt,andtheRed is
Sea 1.2Mt. Dust particles bring nutrients to marine ecosystems, which is especially cu
s
important for the oligothrophic Northern Red Sea. However, their contribution to the s
io
nutrientbalanceintheRedSearemainslargelyunknown.Byscalingtheeffectofone n
P
stormtothenumberofduststormsobservedannuallyovertheRedSea,weroughly a
20 p
e
estimatetheannualdustdepositiontotheRedSeatobe6Mt. r
|
1 Introduction D
is
c
u
Mineraldust is thedominant atmosphericaerosol (Buseckand Posfai,1999). Itplays ss
animportantroleintheEarth’sclimatesystem,althoughthemagnitudeandeventhe ion
signofitsradiativeeffectattheTop-of-the-Atmosphere(TOA)remainuncertain.Most P
25 a
p
e
r
19182 |
airborne dust is generated in desert and semi-desert areas and transported across D
is
local-to-globalscales. c
u
Dust storms lift millions of tons of dust into the atmospheric boundary layer. Such ss
io
largequantitiesofdustcausesevereairpollution,reducevisibility,causeairportshut- n
5 downs,andincreasetrafficandaircraftaccidents(Morales,1979;HagenandWoodruff, Pa
p
1973;MiddletonandChaudhary,1988;Dayanetal.,1991;Yong-SeungandMa-Beong, e
r
1996).Otherenvironmentalimpactsofdustincludereducedsoilfertilityandcropdam-
|
age, reduced solar radiation on the surface and, as a consequence, decreased ef-
D
ficiency of solar devices, damaged telecommunications and mechanical systems, in- is
10 creasedoccurrenceofrespiratorydiseasesandotherimpactsonhumanhealth(Hagen cu
s
andWoodruff,1973;Mitchell,1971;Fryrear,1981;Squires,2007;Jauregui,1989;Liu s
io
andOu,1990;Yong-SeungandMa-Beong,1996;NihlenandLund,1995;Longstreth n
P
etal.,1995;Bennettetal.,2006;Bennionetal.,2007).Dustdepositiontooceans,how- a
p
ever, provides nutrients to ocean surface waters and the seabed (Talbot et al., 1986; er
Swapetal.,1996andZhuetal.,1997).
15 |
Despitecontinuousattemptstoimprovemodelingofdustgenerationprocessesand
D
accumulationofempiricaldata,theamountofatmosphericdustaerosols,theiroptical is
c
properties,andtheirradiativeimpactremainuncertain.Inthe1970s,thetotalAeolian u
s
dustemissionsfromdesertswereestimatedatapproximately500Mtyr−1 level(Peter- sio
n
20 sonandJunge,1971).Junge(1979),GanorandMamane(1982),andMorales(1979) P
found that the Sahara desert generates 200 to 330Mtyr−1 or about 50% of the to- ap
e
tal annual dust flux. These studies appear to underestimate dust emissions, however r
recentcalculations(Andrea,1995;Duce,1995)showthatdustemissionstotheatmo- |
sphereonaglobalscalearetypicallyintherangeof1000to3000Mtyr−1.Tanakaand
D
Chiba (2006) calculated that the annual dust emission from North Africa is 1087Mt is
25 c
u
andfromAsia(includingtheArabianPeninsula,CentralAsiaandChina)575Mt.Inter- s
s
governmentalPanelonClimateChange(IPCC,2001)reportedthemineraldustmass io
n
flux from Asia to be about 100 to 200Mtyr−1, which is approximately 10% of total P
a
annual global dust emissions. IPCC (2007) later reported that Asian dust emissions p
e
r
19183 |
increased to about 800Mtyr−1 with the Taklamakan and Gobi Deserts being the two Dis
majorsourcesofAsiandustemissions(Unoetal.,2005)dispersingdustontheglobal cu
s
scale(Clarkeetal.,2001;Groussetetal.,2003). s
io
Direct and indirect atmospheric radiative impact by dust has implications on global n
P
climate change and presently is one of the largest unknowns in climate model pre- a
5 p
dictions. The African and Asian low-latitude deserts, the so-called dust belt, are the er
major sources of dust for the entire world (Pye, 1987; Prospero et al., 1987, Arimoto
|
etal.,1989;Laurentetal.,2008;Unoetal.,2005).ThedustbeltincludestheSahara
D
desert,aridandsemiaridregionsinArabiaandCentralAsia,andtheTaklamakanand is
c
Gobi deserts in East Asia. Littmann (1991), Wang et al. (1996), Xu and Hu (1997), u
10 s
s
Zhou(2001),Qiuetal.(2001)andQianetal.(2002)studiedspatialandtemporaldust io
n
variability,sourceareasandmovingpathsofduststormsinChina.Duringseveredust P
a
storms, dust from East Asia can reach far beyond the continent, drifting over the Pa- p
e
cific Ocean to the west coast of North America (Husar et al., 2001; Tratt et al., 2001; r
15 McKendryetal.,2001).Similarly,SaharandustcrossestheAtlanticandaffectsboth |
Americas(Prosperoetal.,1987;ReidandMaring,2003;Haywoodetal.,2003). D
The Sahara is the world’s largest dust source (Washington et al., 2003; Shao is
c
u
et al., 2011). The characteristics of Saharan dust were studied experimentally since s
s
the 1990s (Jayaraman et al., 1998; Satheesh and Ramanathan, 2000; Haywood and io
n
Boucher, 2000). The Puerto Rico Dust Experiment (PRIDE) (Reid and Maring, 2003) P
20 a
andtheSaharanDustExperiment(SHADE)(Haywoodetal.,2003)focusedonSaha- p
e
r
ran dust during long-range transport. Two comprehensive field campaigns were con-
ductedin2006(Heintzenberg,2009)and2008(Ansmannetal.,2011)undertheframe- |
work of the Saharan Mineral Dust Experiment (SAMUM) to quantify the optical prop- D
is
25 erties and the radiative impact of Saharan dust near source regions and in the aged cu
mixedplumefromSaharandustandbiomassburningaerosolstransportedofftheWest ss
Africancontinent.TheFennec2011campaign(Washingtonetal.,2012)madesignifi- ion
cantadvancesbyprovidingnewmeasurementsclosetodustsourcesovertheremote Pa
p
Sahara. e
r
19184 |
Miller and Tegen (1998) conducted one of the first studies of the radiative effects D
is
ofdustusingaGlobalGeneralCirculationModel(GCM)coupledwithamixedocean c
u
module to show that dust aerosols reduce the global average surface net radiation ss
by nearly 3Wm−2 during summer in the Northern Hemisphere. Tegen et al. (1996) ion
P
5 showed that dust from disturbed soils causes a decrease in the net surface radiation a
byabout1Wm−2.Recently,Balkanskietal.(2007)confirmedthattheglobalaverage pe
r
dust-related surface net radiative flux change is in the range of −1.11 to 0.92Wm−2.
However,theradiativeeffectofdustindesertregionsduringdustoutbreakscouldbe |
D
twoordersofmagnitudestronger(Kalenderskietal.,2013). is
10 Radiative effect of dust is quite sensitive to aerosol composition and the refractive cus
index,surfacereflectivity,aerosolsizedistributionandtheverticalprofileofdustparti- s
io
cles.Earlystudiesoftheaerosoleffectsofdustfocusedontherefractiveindex.Sokolik n
P
andToon(1999)andClaquinetal.(1999)showedthatdustmineralogyiscomprised a
p
ofsixmainminerals:illite,montmorillonite,kaolinite,quartz,calciteandhematite.Ab- er
sorption by dust in the shortwave spectrum is mostly determined by the presence of
15 |
ironoxidesintheformofhematite.Balkanskietal.(2007)proposedthatuncertainties
D
be constrained based on the hematite concentration. By considering three hematite is
c
concentrations (0.9, 1.5 and 2.7%), they were able to achieve good agreement with u
s
s
Aerosol Robotic Network (AERONET) retrieved refractive indices and show that dust io
n
20 islessabsorbingthanwaspreviouslythought. P
Zhang et al. (2013) showed that the radiative impact of dust vary for the different ap
e
dustverticalprofilesovertheAfricanandAsiandustsourceregionsinsolarspectrum r
byabout0.2to0.25Wm−2 andinInfraredbyabout0.1to0.2Wm−2.Toquantifythe |
vertical distribution, transport, generation, and deposition of dust, global off-line dust D
modelsweredevelopedbyZenderetal.(2003),Chinetal.(2002),Ginouxetal.(2004), is
25 c
u
andTanakaandChiba(2006).Modelstudiesshowedthatabout30%ofemitteddust s
s
is deposited very near the dust sources and 70% is subject to regional and long- io
n
rangetransport.Prospero(1996)estimatedthatdustdepositionintheAtlanticisabout P
a
170Mtyr−1,25Mtyr−1 intheMediterranean,and5Mtyr−1 intheCaribbean. p
e
r
19185 |
Ithaslongbeenrecognizedthatradiativeeffectsofmineralaerosolsareanimportant D
is
driveroftheclimate(TegenandFung,1994,1995;Tegenetal.,1996;Lieetal.,1996; c
u
Andreas, 1996). Early attempts at dust modeling using global models were made by ss
io
Westphal et al. (1987, 1988), Gillette and Hanson (1989), Joussaume (1990), Tegen n
5 and Fung (1994, 1995). The presence of dust affects atmospheric heating rates and Pa
p
thus atmospheric stability and circulation, which together with changes in the surface e
r
energybalanceaffectthehydrologiccycle(MillerandTegen,1998;Milleretal.,2004b).
|
The climate is altered by the presence of dust and the altered climate can in turn
D
influencetheemission,transport,anddepositionofaerosols(Milleretal.,2004b;Yue is
10 et al., 2010). These dust effects on a global scale generally cause a reduction in the cus
atmospheric dust load (Perlwitz et al., 2001). A comprehensive analysis of simulated s
io
dustdistributionsandtheirradiativeeffectswasrecentlyconductedunderthescopeof n
P
theAerosolInter-comparison(AeroCom)project(Textoretal.,2006;Stieretal.,2007). a
p
Regional-scaledusteffectswereaddressedusingfine-resolutionlimitedareamodels er
(e.g., Nickovic et al., 2001; Uno et al., 2001; Lu and Shao, 2001; Kaskaoutis et al.,
15 |
2008;Tegenetal.,2013).Aregionalmodelingsystemwasdevelopedforsimulations
D
ofSaharandustemissions,transportandradiativeeffects(Heinoldetal.,2008,2009; is
c
Laurentetal.,2010;Tegenetal.,2010)duringSAMUM-1.Aregionalmodelwasalso u
s
s
usedtosimulatethespatio-temporalevolutionofthemixedplumeofSaharandustand io
20 biomassburningaerosols(Heinoldetal.,2011a),theradiativeeffect,andthedynamic nP
a
responseoftheatmosphereduetoSaharandustandbiomassburning(Heinoldetal., p
e
2011b)duringSAMUM-2. r
ThestudiesmentionedthusfarmostlyfocusedonwesternandcentralNorthAfrica, |
theSahel,andSahararegions.Therearerelativelyfewermeasurementandmodeling D
studies focusing on eastern North Africa and the Arabian Peninsula. This region has is
25 c
u
been severely under-sampled; there are few available in-situ observations, and very s
s
fewinternationalresearchcampaignshavebeenconductedinthisareacomparedwith io
n
theSahara.IntheArabianPeninsula,duststormsandblowingdustarefrequentevents P
a
duringmostoftheyear.ThemajordustsourcesintheArabianPeninsulaincludethe p
e
r
19186 |
Tigris and Euphrates Rivers, the alluvial plain in Iraq and Kuwait, the low-lying flat D
is
lands in the east of the peninsula along the Persian Gulf and the Ad Dahna and the c
u
Rub Al Khali deserts (Kutiel and Furman, 2003; Shao, 2008). Goudie and Middleton ss
io
(2006) provided an extensive discussion of the causes of dust storms in the Middle n
P
5 East.EasternSyria,northernJordanandwesternIraqarethesourceformostofthe a
p
finedustparticles(lessthan50µmindiameter)foundinArabianduststorms(NOAA, e
r
2002).ThisdustistransportedfarintosouthernArabiawhereitmayhavecontributedto
|
loessdepositsonpartsoftheArabianShieldandevenintheAsirPlateau.Sandstorms
D
containing larger particles (150 to 300µm in diameter) are also frequent throughout is
10 Arabia, but rarely reach above 15m altitude. Most sand grains move by saltation and cus
some by surface creep; they rarely move in suspension by storm winds, haboobs, or s
io
dustdevils. n
P
Dust storms in the southern Arabian Peninsula are more frequent in summer. In a
p
the northern Arabian Peninsula they occur mainly in spring. The peak of dust storm er
activity occurs usually during the day time, when intense solar heating of the ground
15 |
generatesturbulenceandlocalpressuregradients(Middleton,1986a,1986b).Across
D
theArabianPeninsula,remotelysensedaerosolanddustactivitypeaksduringMayto is
c
August (Prospero et al., 2002; Washington et al., 2003; Barkan et al., 2004; Goudie u
s
and Middleton, 2006). The United Arab Emirates Unified Aerosol Experiment (UAE2) sio
n
20 focused on dust in the southern Arabian Gulf region in August to September 2004 P
a
to evaluate the properties of dust particles that converge in the UAE region from nu- p
e
merous sources and their impact on the radiation budget (Reid et al., 2008). Mohalfi r
et al. (1998) studied the effect of dust aerosols on the synoptic system and showed |
that dust aerosol radiative heating strengthens Saudi Arabian heat low. Kalenderski D
etal.(2013)simulatedawinterdusteventthatoccurredinJanuary2009overtheAra- is
25 c
u
bianPeninsulaandtheRedSeatostudyvariousdustphenomena.Notaroetal.(2013) s
s
investigatedthetemporalandspatialcharacteristicsofSaudiArabianduststormsus- io
n
ingtrajectoryanalysis.UsingMODISdata,theyfoundthatthehighestaerosoloptical P
a
p
e
r
19187 |
depth (AOD) is achieved during dust storms that originate from the Rub Al Khali and D
is
IraqiDeserts. c
u
This study aims to quantify the impact of dust on the Arabian Peninsula and Red ss
io
Seafocusingonthesevereduststormobservedduring18–19March2012.Thisstorm n
P
5 likelydrewatleastsomeofitsdustfromIraq,Iran,KuwaitandtheArabianPeninsula’s a
p
Empty Quarter or Rub’ al Khali desert, sprawling over parts of Saudi Arabia, Yemen, e
r
OmanandUnitedArabEmirates.Thesimulationperiodwasfrom00:00UTC(Universal
|
TimeCoordinate)on5Marchto00:00UTCon25March2012withthemodelspin-up
D
during first six days. Using model output and observations, we provide an improved is
10 estimateofthestorm’simpactontheterrestrialandoceanicenvironment. cus
Theremainderofthispaperisorganizedasfollows.Section2discussesthemodel’s s
io
configurationandmethodology.Section3examinestheenvironmentalconditionsthat n
P
ledtodevelopmentoftheduststorm,estimatesthedustemissions,loadanddeposi- a
p
tion,discussesthestructureoftheduststorm,comparesmodelresultswithavailable er
AerosolRoboticNetwork(AERONET)andsatelliteobservations,presentsthevertical
15 |
structure of atmosphere and dust distribution, and calculates radiative effect of dust.
D
WeofferasummaryoftheresultsinSect.4. is
c
u
s
s
io
2 Methodology n
P
a
p
In this study, we combine advanced high-resolution modeling of meteorological and e
r
dust processes with analysis of ground-based and satellite observations of aerosols
20
|
andmeteorologicalfields.
D
is
2.1 Modeling c
u
s
s
The Weather Research Forecast (WRF) system is a mesoscale forecast model with io
n
anincorporateddataassimilationcapabilitythatadvancesboththeunderstandingand P
a
prediction of weather (Skamarock et al., 2008). WRF has been utilized in a variety of p
25 e
r
19188 |
researchandoperationalprojects,fromthescaleofconvectivestormstothescaleof D
is
continental weather patterns (Michalakes et al., 2005). WRF-Chem extends WRF by c
u
incorporatingachemistrymodulethatinteractivelysimulatesemissionsofaerosolsand ss
io
gases,theirtransport,turbulentandconvectivemixing,andchemicalandmicrophysical n
P
5 transformationsoftracegasesandaerosols(Grelletal.,2005). a
p
Here, we configured WRF-Chem with the RADM2 (Regional Acid Deposition e
r
Model 2) photochemical mechanism (Stockwell et al., 1990), the Fast-j photolysis
|
scheme (Wild et al., 2000), the Modal Aerosol Dynamics Model for Europe (MADE)
D
and the Secondary Organic Aerosol Model (SORGAM) aerosol model (Ackermann is
10 etal.,1998;Schelletal.,2001).MADE/SORGAMusesthemodalapproachwiththree cus
log-normally distributed modes (Aitken, accumulation, and coarse modes) to repre- s
io
senttheaerosolsizedistribution.TheaerosolspeciestreatedinMADE/SORGAMare n
P
mainly composed of sulfate, nitrate, ammonium, organic matter (OM), black carbon a
p
(BC),water,seasaltandmineraldust.WealsoemployedtheGoddardGlobalOzone er
ChemistryAerosolRadiationandTransport(GOCART)dustemissionscheme(Ginoux
15 |
etal.,2001)tocalculatetheinfluxofdustintotheatmosphere.GOCARTsimulatesdust
D
emissionsasafunctionofsurfacewindspeed,surfaceerodibility,andsurfacewetness is
c
(Chinetal.,2002).Emissionflux,F ,foraspecificaerosolsizegroup,p,isexpressed u
p s
s
as io
n
P
F =CSs U2 (U −U ) if U >U (1) a
20 p p 10m 10m t 10m t pe
r
where C is a dimensional constant coefficient of proportionality; S is the space-
|
dependent dimensionless erodibility field taken from Ginoux et al. (2001) and has
a spatial resolution of 0.25◦×0.25◦; U10m is the wind speed 10m above the ground; Dis
c
25 Utisthethresholdvelocityofwinderosion,whichdependsonparticlesizeandsurface us
wetness;sp isaparticularparticlesizemassfractionwithinthesoil.Dustisaprimary sio
aerosolanddoesnothaveaveryfineAitkenmode.s issettobe0,0.07and0.93for n
p P
Aitken,accumulationandcoarsemodes,respectively. a
p
e
r
19189 |
Thedustgenerationprocessisverycomplex.Eq.(1)isasimplifiedempiricalapprox- D
is
imationofthisprocessandhastobeadjustedusingavailableobservationstoaccount c
u
for different model resolution and meteorological conditions. For example, coefficient ss
C in Eq. (1) was originally estimated to be equal to 1mgs2m−5 based on regional ion
P
5 datasets(Ginouxetal.,2001).Zhaoetal.(2010)proposedtotunethisparameterand a
usedC=0.65mgs2m−5intheircalculationstoreplicateconsistencyofcalculatedop- pe
r
ticaldepthwithAERONETobservations.Inoursimulations,weusedC=0.8mgs2m−5
|
toachieveconsistencybetweenthesimulatedandAERONETaerosolopticalthickness
D
(AOD)observedfrom11to24March2012. is
10 Tocovertheentireareaaffectedbythestormandtoaccountforprocessesthatwere cus
responsible for the storm’s development, we ran the WRF-Chem model in the spatial s
io
domain from 4 to 40◦N in longitude and 25 to 80◦E in latitude with 10km horizontal n
P
resolution. In our calculations, we therefore used 495 grid points in the west-east di- a
p
rection and 396 grid points in the south-north direction, as well as 40 vertical levels er
with the top of the model domain at the 10hPa level. The following physical parame-
15 |
terizations(Skamarocketal.,2008)wereusedtoconfigurethemodelsimulations:the
D
Linmicrophysicsscheme;theRapidRadiativeTransferModel(RRTMG)forbothlong- is
c
wave (LW) and shortwave (SW) radiation; the Mellor-Yamada-Janjic (MYJ) boundary u
s
s
layerscheme;theNoahlandsurfacemodel;theGrellcumulusparameterization.Lat- io
n
20 eralboundaryandinitialconditionsformeteorologicalfieldswereprovidedbytheNa- P
a
tionalCentersforEnvironmentalPrediction’s(NCEP)globalanalysis(FNL).Weused p
e
NCEP’sdailyglobalseasurfacetemperature(SST)analysis(RTG_SST_HR)toupdate r
SSTeverysixhours.Thevariousmodeldomainconfigurationsandphysicsoptionsare |
summarizedinTable1. D
is
c
2.2 Observations u
25 s
s
io
Totestandconstrainthemodelsimulations,weusedreanalysisfieldsaswellasme- n
P
teorological and aerosol observations available in the Middle East and surrounding a
p
e
areas. r
19190 |
2.2.1 Space-borneaerosolinstrumentalobservations D
is
c
u
AerosolOpticalDepth(AOD)datafromMODIS(ModerateResolutionImagingSpectro- s
s
radiometer)sensorsonboardtheTerraandAquasatellitesarewidelyusedinaerosol io
n
studies (Salomonson et al., 1989). The MODIS instrument provides high radiometric P
a
5 sensitivity(12bit)in36spectralbandsranginginwavelengthfrom0.4µmto14.4µm. pe
r
Two bands are imaged at a nominal resolution of 0.25km at nadir, with five bands at
0.5km, and the remaining 29 bands at 1km. A ±55◦ scanning pattern of the Earth |
ObservingSatellite(EOS)orbitat705kmaltitudeachievesa2330kmswathandpro- D
is
videsglobalcoverageeveryonetotwodays.TheMODISdailyLevel2AODdataare c
u
10 producedatthespatialresolutionofa10km×10km(atnadir).TherearetwoMODIS ssio
aerosol data products, one containing data collected from the Terra platform and the n
P
other containing data collected from the Aqua platform. To maximize the observation a
p
coverage,weusedAODretrievalsoverlandandseaderivedfromthedarktargetprod- e
r
uct (Remer et al., 2005, 2008) and the “deep blue” product over bright land surfaces
|
(Hsu et al., 2004, 2006). The dark target ocean and land AOD products were avail-
15
D
able from both Terra and Aqua, but the deep blue retrievals were only available from is
Aqua.Inthisstudy,wehaveuseddailydeepblue(level2)dataincombinationwiththe cu
s
standardoceanalgorithmforcomparisonwiththesimulatedaerosolopticalproperties. s
io
Spinning Enhanced Visible Infrared Radiometer (SEVIRI) (Aminou et al., 2002) is n
P
alinescanningradiometercurrentlylocatedonboardtheEuropeangeostationaryme- a
20 p
e
teorologicalsatellite,Meteosat-9;itwaspreviouslyonMeteosat-8.Thisinstrumentpro- r
videsdatainfourvisibleandnearinfraredchannelsand8infraredchannelswithares-
|
olution of 3km at nadir. In this study, we used the 0.6µm channel. An advantageous
D
feature of SEVIRI is its ability to image the study area with high temporal resolution is
c
of15min.ThisallowsSEVIRItotrackaerosolevents,whichoffersagreatadvantage u
25 s
s
overpolarorbitinginstruments,likeMODIS,thatusuallyseestheparticularseenonce io
n
aday.TheSEVIRI’sspatialresolutionis,however,coarserthanthatofMODISbutthe P
a
p
e
r
19191 |
aerosolgriddedproductsfortheArabianPeninsulafrombothinstrumentsareroughly D
is
onthe10km×10kmgrid. c
u
s
s
2.2.2 Ground-basedaerosolobservations io
n
P
a
TheAerosolRoboticNetwork(AERONET)usesCIMELRoboticSunPhotometersand p
e
provides observations of AOD on up to 8 wavelength channels between 0.340 and r
5
1.640µm (Holben et al., 1998) and angular distribution of sky radiance at four wave- |
lengths(0.440,0.675,0.870,and1.020µm).Inadditiontothedirectmeasurementof D
AOD,theinversionalgorithmretrievesphysicalandopticalpropertiesofaerosolsthat isc
u
comprisetheaerosolrefractiveindex,columnaveragesizedistribution,singlescatter- s
s
10 ing albedo, and asymmetry parameter. The maximum AERONET uncertainty in AOD ion
retrieval is estimated to be 0.02 with the highest error in the ultraviolet wavelength P
a
p
(Holbenetal.,1998;Ecketal.,1999)andthecalibratedskyradiancemeasurements e
r
typically have an uncertainty of less than 5% (Holben et al., 1998). Cimel Sun Pho-
tometersarecalibratedannuallybycomparisonwithanAERONETmasterinstrument. |
The AERONET data are at high time resolution but have limited spatial cover- D
15 is
age. During the study period, AOD observations were available at Kuwait Univer- c
u
sity (29.32◦N, 47.97◦E), at Mezaira (23.14◦N, 53.77◦E), and at our own site estab- ss
lishedinFebruary2012ontheKAUSTcampus(22.30◦N,39.10◦E).Wechosecloud ion
P
screened(Level1.5)AERONETAODforourmodelvalidationsincelevel2data(cloud a
p
screened and quality assured data) were not available during the simulation period. e
20 r
TheAngstrompowerlawwasusedforcomparingthesimulatedopticalpropertiesout-
|
putat0.60µmwiththeAERONETmeasurementsasfollows:
D
(cid:18)0.600(cid:19)−α isc
AOD(0.600)= AOD(0.675)× , (2) us
0.675 s
io
n
P
a
p
e
r
19192 |
whereαistheAngstromexponentcalculatedfromtheAERONETmeasurementas: D
is
c
h i u
ln AOD(0.440) ss
α= lnA(cid:0)O0D.(607.657(cid:1)5) (3) ionP
0.440 a
p
e
2.2.3 Weatherstationmeteorologicalobservations r
|
Therearemeteorologicaldatafrom108surfaceweatherstationsand12upperairsta-
D
5 tions available in the area of interest. For the purposes of this study, we focus on the is
c
upper air radiosonde data. The radiosonde (RS) is a balloon-borne instrument plat- u
s
form with radio transmitting capabilities. Data from the radiosonde is interpreted at sio
thelaunchingstationandenteredintoaworldwidecommunicationsnetwork.Thetime n
P
sampling of RS data is usually twice daily at 00:00UTC (midnight) and 12:00UTC ap
e
10 (noon). RS profiles contain pressure, temperature, relative humidity, wind speed, and r
winddirection.Dependingonthesensorandballoonusedateachlocation,thevertical |
resolutionofRSdatatypicallyhave40levelsbetween1000and50hPa.Weusedtem-
D
perature profiles from three RS stations for a comparison with modeled heating rates is
c
at12:00UTCon19March2012.StationsareidentifiedbytheirWorldMeteorological us
s
15 Organization(WMO)fivedigitcodes.ThethreestationsovertheArabianPeninsulase- ion
lectedforthisstudyincludeJeddah(41024),AbuDhabi(41217)andKuwait(40582). P
a
The vertical temperature RS profiles were retrieved from the upper air archive at the p
e
UniversityofWyoming(http://weather.uwyo.edu/upperair/sounding.html). r
|
2.2.4 Reanalysisoutput
D
is
c
To compare meteorological fields and spatial patterns, we use output from the ERA- u
20 s
Interim reanalysis (Dee et al., 2011). It is the latest global atmospheric reanalysis sio
n
produced by the European Centre for Medium-Range Weather Forecast (ECMWF)
P
as a transition between ERA-40 and a future reanalysis project. It provides infor- ap
e
mation on a large variety of surface parameters (3hourly), describing weather as r
19193 |
well as ocean-wave and land-surface conditions and 6hourly upper-air parameters D
(37 pressure levels up to 1hPa), on a 0.7◦×0.7◦ grid. ERA-Interim uses improved isc
u
an atmospheric model and a more sophisticated data assimilation method (4-D-Var) ss
io
for atmospheric analysis compared with ERA-40. Information about the current sta- n
P
5 tus of ERA-Interim production and availability of online data can be found at (http: a
p
//apps.ecmwf.int/datasets/data/interim_full_daily/). In this paper, 10m wind field (U ) e
10 r
from WRF-Chem high resolution (10km) simulation results are compared with ERA-
|
Interimreanalysis.TheU fieldfromERA-InterimwasinterpolatedtoWRFprojections
10
D
forcomparison. is
c
u
s
s
10 3 Results ion
P
a
Theduststormon18–22March2012coveredanextremelylargeareaanddisrupted p
e
human activities in Iraq, Iran, Kuwait, Syria, Jordan, Israel, Lebanon, UAE, Qatar, r
Bahrain,SaudiArabia,Oman,Yemen,Sudan,Egypt,Afghanistan,andPakistan.Many |
airports in these countries were shut down because of dust impact on visibility and D
15 machinery. iscu
s
s
3.1 Meteorologicalconditionsandstormspatial-temporaldevelopment io
n
P
a
Figure1showsgeopotentialheightfromWRF-ChemsimulationsandERA-Interimre- p
e
analysis at 500hPa at 00:00UTC on 17–19 March 2012. Sea-level pressure for the r
samedaysat00:00UTCisshowninFig.2.Themodelresultscomparewellwiththe |
20 reanalysisfieldsinbothFigures.On17Marchtherewasatroughinmid-troposphere D
at 500hPa stretching over eastern Mediterranean Sea, Syria, Iraq, western Iran and is
c
u
northern Saudi Arabia. This upper air disturbance caused a significant pressure de- s
s
crease in the lower atmosphere and a low-pressure system consequently developed io
n
in eastern Iraq and northwestern Iran with a central sea-level pressure of 1008hPa. P
a
A high pressure is seen over northeastern Afghanistan and Tajikistan. The spatial p
25 e
r
19194 |
distribution of 10m wind bars at 09:00UTC from 17 to 19 March is shown in Fig. 3. D
Strongsurfacewindsof15ms−1 areobservedovernorthwesternIraqon17Marchat isc
u
09:00UTC(Fig.3a). ss
io
On 18 March the high level trough deepened and moved southeastward into the n
P
5 northern Arabian Peninsula covering Iraq and western Iran. A surface high pressure a
p
builds up over the eastern Mediterranean Sea. The presence of a low pressure in e
r
southernSaudiArabia,theGulfofOmanandcentralIranincombinationwithahighin
|
theeasternMediterraneanSeaenhancedastrongpressuregradientalongtheArabian
Peninsula.Asaresult,astrongsurfacewindof15ms−1 developedtothewestofthe Dis
10 lownearthePersianGulfinsoutheasternSaudiArabiaandQatarandtothenortheast cus
of UAE (Fig. 3b). This is well above the generally assumed threshold wind speed (5 s
io
to 6ms−1) for dust suspension (Gillette, 1978). These strong winds forced the dust n
P
emissionandtransporteddusttothesouthandsouthwestoftheArabianPeninsula. a
p
On 19 March the high level trough moved towards eastern Iran and western er
Afghanistan. The low-pressure system in the southern Arabian Peninsula was over-
15 |
takenbythehigh-pressuresystem.ButifasurfacehighdevelopedinthenorthernAra-
D
bianPeninsula,itcouldcausefurtherdusttransporttowardsthesouthandsouthwest is
c
oftheArabianPeninsula.Alow-pressuresystemthatdevelopedincentralPakistanin u
s
s
combination with the high pressure in the northern Arabian Peninsula increased the io
20 strong winds in southern Iran with a strong surface wind speed of 20ms−1 (Fig. 3c). nP
a
OneofthemostidentifiablesynopticfeaturesintheArabianPeninsuladuringthesim- p
e
ulation is the low-pressure system that moved southeastward through the Peninsula r
on 19 March 2012 suggesting that the dust storm was closely related to the south |
and southwestward propagation of a cold front. Figure 4 shows the simulated near- D
surface wind field and dust concentration at 08:00UTC and 15:00UTC on 18 March is
25 c
u
2012.Comparisonswithsynopticobservationsconfirmthattheflowfieldiswellsimu- s
s
lated.Thecoldfrontisaline-shapednarrowregionstretchingfromthewesterncoast io
n
of Saudi Arabia to Eastern UAE, as can be seen from the dense temperature con- P
a
tours (Fig. 4a). Strong northeasterly wind (maximum 15ms−1) prevailed behind the p
e
r
19195 |
cold front, resulting in strong dust emissions and high dust concentrations. The dust D
is
wascarriedsouthwestwardbythenortheasterlywindtoconvergeonthefrontalarea, c
u
resulting in high dust concentrations near the front. The dust was then transported ss
io
westward over the red sea along the front and towards the southern Arabian Penin- n
5 sula. The dust storm affected a large area, although the area of dust emissions was Pa
p
much smaller. In southern Saudi Arabia, the dust front rapidly advanced to the south e
r
andsouthwest.Figure4showsthatwithinatimeperiodofsevenhours,thedustfront
|
advancedmorethan150km.At15:00UTC18March(Fig.4b),dustwaswidespread,
D
withthesamemaximumdustconcentrationasthatat08:00UTC. is
10 Dust and sand lift both ahead of and behind cold fronts; winds tend to be even cus
strongerbehindthefrontthanaheadofit(Sissakianetal.,2013).Behindanadvanc- s
io
ing cold front, southern Iraq and northern Saudi Arabia experienced blowing dust or n
P
shamal-likeconditions(Hamidietal.,2013).Thesummershamalresultsfromacharac- a
p
teristicsynopticpatterninwhichthemostprominentfeaturesare(a)asemi-permanent er
high-pressurecellextendingfromtheeasternMediterraneantonorthernSaudiArabia;
15 |
(b)alow-pressurecelloverAfghanistan;and(c)athermallowpressurebeltassociated
D
withthemonsoontroughextendingintosouthernArabianpeninsula.Ahighpressure is
c
intheeasternMediterraneanandalowpressureovertheAfghanistanareclearlyseen. u
s
s
io
3.2 Theverticallyintegratedmassbalanceandopticaldepthofthedust n
P
a
p
Dust emissions are calculated in the model interactively from Eq. (1) using soil prop- e
20 r
erties and simulated surface winds that drive dust emissions and are extremely im-
|
portant to storm development. Therefore, to ensure that the model could adequately
D
generate dust, we compare thesimulated 10m wind field (U10m) with the output from is
c
theECMWF(EuropeanCenterforMediumRangeWeatherForecast)Interimglobalre- u
s
25 analysis,ERA-I.TheU10misextractedfromERA-IandinterpolatedtoWRFprojections sio
forcomparison.Thesimulatedandreanalysisfieldsarewellcorrelated(notshown)with n
P
a time-averaged correlation coefficient of 0.79. Figure 5a shows the erodibility field S a
p
e
used in the GOCART emission scheme (Eq. 1). Figure 5b–d shows simulated dust r
19196 |
emissionsaveragedover24hperiodsfor17,18and19March2012,respectively.On D
is
17 March (Fig. 5b) the most intense simulated emissions were located in the lower c
u
TigrisandEuphratesrivervalleysandinnewlydrylakebeds,whicharefilledwithloose ss
io
fine-grainsoilsandhavehighsusceptibility.Theseareknownmajorsourcesofdustin n
5 Iraq.Themaximumdustemissionrateexceeded300µgm−2s−1.On18March(Fig.5c) Pa
p
dustwasgeneratedinthewideArabianDesertarea(TheRub’alKhali,AnNafudand e
r
AdDahna)withalmostsamemaximumintensityasonthepreviousday.Adustsource
belt about 50km wide along the west coast (22◦N to 25◦N) of the Arabian Peninsula |
D
was also active on 18 March. On 19 March (Fig. 5d), the dust source regions were is
10 alongtheeasterncoastofthePersianGulf,coastalareasoftheArabianSeanearIran cus
andPakistan,andonthesouthernflanksofthemountainchainalongtheArabianSea s
io
coast.Theactivedustsourceinthisregionwasasmallintermountainvalleycentered n
P
at 27.5◦N, 59◦E. At the center of this valley, there is a large, 1087km2, dry salt lake, a
p
Hamum-e-JazMurian.ThemajorityofdustattheeasterncoastofthePersianGulfwas er
generatedfromthislake.DustsourceregionsalongthewesterncoastoftheArabian
15 |
Peninsula were also active. The area of dust generation was relatively compact but
D
maximumdustemissionsexceeded500µgm−2s−1 duetostrongwinds(Fig.3). is
c
Thespatialdistributionofdailyaveragesimulateddustdepositionratesfor17,18and u
s
s
19 March 2012 are shown in Fig. 6. There was no precipitation during the simulation io
n
20 period and therefore the dominant dust removal mechanism was dry deposition. The P
a
deposition velocity is higher for large particles therefore coarse dust tends to deposit p
e
closer to a source than the fine ones. Figure 6 shows that the maximum dust depo- r
sition exceeds 90µgm−2s−1, about one third of maximum dust emissions rates. On |
17 March (Fig. 6a) a lot of dust was deposited in northeastern Persian Gulf relatively D
far from the source regions. On 18 March the strongest dust deposition happened in is
25 c
u
widesurroundingsofthedustsources,aswellasinthePersianGulfandtheRedSea s
s
(Fig.6b).On19March(Fig.6c),dustwasdepositedintheRedSea,theGulfofOman io
n
andtheArabianSeaofftheIranandPakistancoasts. P
a
p
e
r
19197 |
Totaldustemission/depositionarecomputedbyintegratingemission/depositionover D
is
thedomainarea.Figure7ashowstimeseriesofdailytotalemissionanddeposition(in c
u
Mtday−1) for the simulation period of 11 to 24 March 2012. The daily emissions (Fs) ssio
anddepositions(F )areofthesameorderofmagnitude.Themaximumdailyemission, n
d
5 Fs=13.10Mtday−1, and maximum daily deposition, Fd=10.06Mtday−1, occurred on Pap
18 March because dust was dispersed and deposited over wider areas than those e
r
where dust was generated. The maximum total hourly dust emission and deposition
|
rates of 0.85Mthr−1 and 0.52Mthr−1, respectively, occurred at about 07:00UTC on
D
18 March (not shown). This is expected because in the daytime the atmosphere is is
c
unstable and the momentum transfer from the free troposphere to the surface, which u
10 s
enhancedthesurfacewinds,ismoreefficient.Dustemissionsaresubstantiallyweaker sio
inthenightwhentheboundarylayercollapses.Zhaoetal.(2010)simulatedthetotal n
P
dust emissions using DUSTRAN and GOCART dust schemes over Northern Africa ap
using WRF-Chem and found maximum daily dust emissions of 12Mtday−1, which is er
15 similar in magnitude to the emission flux calculated in this study. Shao et al. (2003) |
estimatedaveragedailytotaldustemissionsandtotaldustdepositioninaNortheastern
D
Asian Dust storm that occurred in 2002. They reported the average emissions and is
c
deposition to be 11.5 and 10.8Mtday−1, respectively. In our study, the average daily us
total dust emission and deposition were 6.7 and 5.2Mtday−1, which is about half of sio
n
20 thatestimatedbyShaoetal.(2003).Thiscomparisonthusshowsthattheduststorm P
a
of 18–22 March 2012 is in the range of the most powerful dust events considered in p
e
theliterature. r
A portion of the dust deposited in the ocean provides nutrients to marine ecosys- |
temsandincreasestheocean’snetproductivity.DustemittedfromsouthernIraq,Iran D
25 and the eastern Arabian Peninsula was mainly deposited in the Persian Gulf and the isc
u
ArabianSea.DustemittedfromthewesterncoastoftheArabianPeninsulawasmostly s
s
depositedinRedSea.Timeseriesofdailytotaldustdepositionovertheoceanareas io
n
(including the Red Sea, Persian Gulf, and parts of the Arabian Sea that are included P
a
in the domain) are plotted in Fig. 7b. Over the ocean areas, the dust load increased pe
r
19198 |
in the course of the dust episode from 17 to 20 March, reaching its maximum on 20 D
is
March(notshown).Thisdelayintimeisexpected,asdustwastransportedfromlandto c
u
ocean.Themaximumdustdepositionof1.5Mtday−1 overoceanareasalsooccurred ss
on 20 March. Over the Red Sea, the maximum dust deposition of 0.23Mtday−1 oc- ion
P
5 curredon19March.Thedustbalanceovertheentiresimulationperiodissummarized a
p
inTable2.Overthe11daysimulationperiod,about93.76Mtofdustwereemittedand e
r
about73.04Mtofdustweredepositedovertheentirestudydomain.Thetotaldustde-
|
positedonthelandwasabout66.48Mtandovertheoceanaround6.56Mt,outofwhich
D
1.20MtweredepositedontheRedSea.Shaoetal.(2007)estimatedthetotalamount is
10 of dust eroded from the Australian continent during the 22–23 October 2002 severe cus
duststormtobe95.8Mt,whichiscomparablewithourpresentvalueof93.76Mtdur- s
io
ingtheMarch2012duststormovertheArabianPeninsula.Zhaoetal.(2006)obtained n
P
a similar figure of 120Mt for the spring dust aerosol emissions into the atmosphere a
p
fromtheAsiansourceregion. er
Althoughthemagnitudeofdustdepositionovertheoceansisnotaccuratelyknown,
15 |
thereisevidencethatdustdepositioncouldhaveasignificantimpactonchemicaland
D
biologicalprocessesintheoceans(Martin,1990;Watsonetal.,2000;Fanetal.2006; is
c
Sunda and Huntsman, 1997). In the present study, we have calculated the amount u
s
s
of dust deposited over the ocean and particularly in the Red Sea. Aeolian deposition io
n
20 is especially important for the Red Sea because this sea has very little fresh water P
a
dischargefromthecoastalareas;whatlittledischargethereismostlyassociatedwith p
e
flashfloodevents. r
DuststormsoccurveryfrequentlyinArabianPeninsulaandtheirintensitycouldbe |
quitedifferentfromyeartoyear.Toestimatethenumberofduststormshappeningan- D
nually,weanalyzedMODISsatelliteimagesforthe2002–2013period,whichwerepro- is
25 c
u
videdunderthescopeofnaturalhazardreportsinthisareabyNASA’srapidresponse s
s
system(http://earthobservatory.nasa.gov/NaturalHazards).Forthis12yearperiod,the io
n
numberofcaseswhenadustplumecoveredmorethan20%oftheMiddleEastarea P
a
reaches237orabout20annually.ThenumberofduststormsaffectedtheRedSeais p
e
r
19199 |
71orabout5–6annually.ThenthedepositionofmineraldusttotheRedSeashouldbe D
about 6Mtyr−1. This figure is a rough estimate. An extended analysis of the intensity isc
u
andvariabilityoftheduststormactivityovertheArabianPeninsulaisrequiredtomake ss
io
thisestimatemorecertain. n
P
5 Figure 8a shows the simulated spatial pattern of the dust load at 10:00UTC on a
p
19March.DustplumeswereseenspreadingovertheGulfofOman,theRedSeaand e
r
thesoutherncoastoftheArabianPeninsula.Thehighestdustloadexceeded4.5gm−2
|
overtheGulfofOmanandsouthernSaudiArabiaborderingYemen.Theimagetaken
D
by Aqua/MODIS at 09:50–10:05UTC (Fig. 8b) shows a significant dust load over the is
10 entire southern Arabian Peninsula and its coastline along the Arabian Sea as well as cus
overthePersianGulf.ThesimulatedandobserveddustloadsinFig.8arequitecon- s
io
sistentatmostlocations,suggestingthatthemodelcorrectlypredictedthedustevent. n
P
Both modeled and satellite imagery clearly show dust aerosols covering not only the a
p
landbutalsoverylargeareasoftheoceans. er
Figure9a–cshowsaspatialdistributionofAODfromWRF-Chem,MODISandSEV-
15 |
ERI retrievals, respectively. Since the number of passes of the MODIS instrument on
D
boardtheAquasatelliteoverthedomainofinterestduringthetimeofinterestislimited, is
c
two0.55µmretrievalsat09:50and10:05UTCon19Marchwereusedforcomparisons u
s
s
withthesimulatedAOD(0.60µm)at10:00UTC.Acombinationofbothstandard-ocean io
n
20 anddeepblueproductswasusedtogetthemaximumspatialcoverageoverthesimu- P
a
lateddomain.ThemodelcapturesthespatialdistributionsofAODwell.Bothmodeled p
e
and observed optical fields indicate a high dust plume over the Persian Gulf, Gulf of r
Oman,RedSeaandsouthernArabianPeninsula. |
Figure 9d–f compares hourly simulated and observed AODs at three AERONET D
sites (KAUST campus, Kuwait University and Mezaira) during the simulation period. is
25 c
u
The AERONET 15min observations were averaged hourly to compare with modeled s
s
values.Thetimeseriesmatchwellwitheachothermostofthetime.Eventhoughthe io
n
magnitudeofthesimulatedAODattheKAUSTcampuswaslowercomparedwiththat P
a
p
e
r
19200 |