DOC

Validation

By Stephen Garcia,2014-02-19 06:55
9 views 0
Validationvalid

    Validation

    ADVANCESINATMoSPHERICSCIENCES,VOL.27,N0.5,2010,11511168

    ValidationofECMWFandNCEPNCAR

    ReanalysisDatainAntarctica

    YULejiang(于乐江),ZHANGZhanhai(张占海),

    DonaldLENSCHOW..HsiaomingHSU..WU

    ZHOUMingyu(周明煜),ShiyuanZHONG,

    Huiding(吴辉碇),andSUNBo(孙波)

    PolarResearchInstituteofChina,Shanghai200136

    .DepartmentofGeography,MichiganStateUniversity,EastLansing,MI48824111,USA

    .NationalCenterforAtmosphericResearch,Boulder,CO80307,USA NationalMarineEnvironmentalForecastCenter,Beijing100081 (Received8September2009;revised18December2009) ABSTRACT

    TheEuropeanCenterforMediumRangeWeatherForecastfECMWF)ReAnalysis(ERA40)andthe

    NationaICentersforEnvironmentalPrediction

    NationalCenterforAtmosphericResearchfNCEPNCAR)

    ECMWFfERA401andNCEP

    NCARreanalysisdatawerecomparedwithAntarcticstationobserva- tions.includingsurface

    layerandupper-layeratmosphericobservations.onintraseasonalandinterannual

    timescales.Attheinterannualtimescale.atmosphericpressureatdifierentheightlevelsinthe

    ERA40data

    areinbetteragreementwithobservedpressurethanthatintheNCEP

    NCARreanalysisdata.ERA40re

    analysisalsooutperformsNCEPNCARreanalysisinatmospherictemperature.exceptinthe

surface1ayer

    wherethebiasesaresomewhatlarger.Thewindvelocityfieldsinbothdatasetsdonotagreewel

    lwith

    surfaceandupper

    layeratmosphericobservations.Atintraseasonaltimescales.bothdatasetscapturethe

    observedintraseasonalvariabilityinpressureandtemperatureduringaustralwinter.

    Keywords:reanalysisdata,interannualvariability,intraseasonalvariability,surfacelayer,u

    pperlayer

    Antarctica

    Citation:Yu,L.,Z.Zhang,M.Zhou,S.Zhong, ofECMWFandNCEP——NCARreanalysisdata

    10.1007/80037601091401.

    1.Introduction

    D.Lenschow,H.Hsu,H.WujandB.Sun,2010:Validation inAntarctica.Adv.Atmos.Sci.,2r(5),11511168,doi:

    MeteorologicalobservationsinAntarcticaare sparse,especiallyintheinteriorofthecontinent.Stud

    iesoftheweatherandclimateoftheAntarcticoften relyupontheoutputsfromvariousglobalorregional atmosphericnumericalmodelsandgriddedglobalre analysisdatasetsderivedfrombothmodelsandob

    servations,suchastheECMWFdatasetfERA.40) (Uppalaeta1.,2005)orthatoftheNCEPNCAR

    EuropeanCenterforMediumRangeWeatherFore-

    cast(ECMWF)ReAnalysisfERA40)(Uppalaeta1.,

    2005)andtheNationalCentersforEflvironmental Prediction(NCEP)NationalCenterforAtmospheric Research(NCAR)globalreanalysis(Kalnayetal 1996;Kistlereta1..2001).Giventheimportanceof thesegriddeddatasetsinunderstandingtheweather

    andclimateoverAntarcticaitisnecessarytoverify whethertheydoinfactrepresentthetruestateofthe atmosphereoverAntarctica,ratherthansimplythe idiosvncraciesofthemodelsusedtoproducethem. Anumberofstudieshaveevaluatedthesereanaly- sisdatasetsusingsurfaceandupperleve1observations

    inAntarctica.UsingobservationsfromtheAntarctic FirstRegionalObservingStudyoftheTroposphere (FROST)in19941995,BromwichandSmithf1993), Turnereta1.(1996),andTurnereta1.(1999)found thattheNCEPNCARandECMWFreanalysisare

    abletoproducethemaincirculationfeaturesover Antarctica.despitesomeminordiscrepanciesinthe Correspondingauthor:ZHOUMingyu,mingyuzhou@yahoo.corn ?

    ChinaNationalCommitteeforInternationalAssociationofMeteorologyandAtmosphericS

    ciences(1AMAS),InstituteofAtmospheric Physics(1AP)andSciencePressandSpringerVerlagBerlinHeidelberg2010 1152VALIDATIONOFREANALySISDATAINANTARCTICA depthsandlocationsofthelows.Cullathereta1. (1997)comparedthetworeanalysisproductsfora10

    yr(19851995)periodandfoundthatsurfacepressure inthetworeanalysisdatasetsisinbetteragreement withobservationsthansurfaceairtemperatureand windandthattheerrorsinbothreanalysisatthe standardpressurelevelsdecreasewithtime.Bromwich eta1.f19951foundthattheECMWFreanalysisdata aresuperiortothoseoftheNCEPNCARinreproduc

    ingtheatmosphericmoisturebudgetinhighsouthern 1atitudesfortheperiod19851992.Bromwicheta1.

    (1999)verifiedNCEPreanalysisdataagainstFROST dataandshowedthatthereanalysisproduceswarmer surfacetemperaturesandweakernear--surfacetemper.. atureinversions.ConnolleyandHarangozof20011also foundpooragreementbetweenobservednearsurface

    temperaturesandthosefromboththeNCEPand ECMWFreanalysis.Kingf2003)usedsurfacepres

    suredataderivedfrominstrumentsdeployedonice floesinsouthernBe11ingshausentoassessECMWF reanalysisdataforFebruaryMa_v2001.Kingfound

    thattheanalysesagreewellwithobservedpressure andpressuregradients.Hineseta1.(2000)andMar shallf20021notedsomespurioustrendsinairtemper

    atureandgeopotentialheightatdifferentlevelsinthe NCEPNCARreanalysisdata.Inseveralstudieson theAntarcticOscillationandEastAsianclimatefFan andng2004;Suneta1.2008;Yueandng

    2008;Zhu,20091,theNCEPNCARreanalysisdata

    werecomparedagainstthoseoftheECMWFreanaly

    sisdataandgoodagreementwasdemonstrated. Thesepreviousstudieshavehelpedthoseworking inthisfieldtounderstandthestrengthsandweak

    nessesofglobalreanalysisproductsoverAntarctica, andthereforetheirusefulnessinimprovingourunder

    standingofweatherandclimateinthisimportantand yetdata-sparseregion.Mostofthesestudies,however havelimitedtheircomparisonstoashorttimeperiod orthesurfaceonly.orasingleheightpressurelevel aloft,andstandardstatistics.

InthisstudytheERA4OandNCEPNCAR(II1

    reanalysisdatasetswereevaluatedusingsurfacelayer

    andupperlayerobservationaldataovera29yrpe

    riodf197920081.Theevaluationswerenotlimitedto surface.1ayervariables,butvariablesatmultipleheight pressurelevelsaloftinordertounderstandwhetherthe reanalysisdatasetscapturetheverticalstructureand variation.Purtherinadditiontousingstandardcorn. parisonstatistics,spectralpropertiesandmajorEm. piricalOrthogonalFunctionfEOF)modeshavealso beencomparedinordertounderstandtheabilityof thereanalysisdataincapturinginterannualandin

    traseasonalvariabilityandthespatialpatternsofthe variability.

    V0L.27

    ThispaperisorganizedaSfollows.Dataandthe analysismethodsaredescribedinsection2.Theval

    idationresultsofthetworeanalysisdatasetsarepre

    sentedinsection3,followedbydiscussioninsection4. Conclusionsaredrawninsection5.

    2.Dataandmethodology

    2.1Reanalysisdatasets

    Thisstudyevaluatestwoglobaldatasets:the NCEPNCARglobalreanalysisandtheERA.40re_ analysis.These1ongterm.dynamicallyconsistent

    globaldatasetsareproducedusingagloba1dataas

    similationsystemwithinputsfromatmosphericglobal modeloutputsandvariousobservationsfrommultiple sources,includingradiosondes,satellites,ships,and aircraft.Thedatasetsconsistofalargesetofatmo-

    sphericandhydrologicalvariablesatthesurfaceand multiplelevelsthroughouttheatmosphere.atmultiple timesperdayonaglobalgrid.

    TheNCEPNCARglobalreanalysisisbasedon

    theNCEPoperationalEtamodelof10January1995 withareducedhorizonta1resolutionofT42f209km) and28verticallevels.Itstemporalcoverageisfour timesperdayfrom1January1948tothepresentday forthefirstreanalysisfNCEPI1andforNCEPIIfrom 1January1979tothepresentdayforthesecondre

    analysisfNCEPII1.Satellitesounderdatawerefirst assimilatedintotheNCEPNCARreanalysisinMarch

    1975(Jenne,2000).From1979onwards,datafromthe TelevisionInfraredObservationalSatellite(TIR0S) OperationalVertica1Sounder0VS)datastartedto

    constraintheNCEPreanalysis(KistlerandKalnay, 20001.Apotentialproblemforthecurrentstudy wasthattheAustralianSurfacePressureBogusData fortheSouthernHemispheresouthof40.Swereread witha180.errorinlongitudebetween1979and1992 fKistlereta1.,20011.Thiserrorhaditsgreatesteffect between45.and60.Sandtheeffectdecreasestoward theSouthPole.Itispossiblethatthiswouldaffectthe comparisonsatintraseasonaltimescales.Detailedde. scriptionsoftheNCEP——NCARglobalreanalysisdata

    aregivenbyKalnayeta1.(1996),Kistlereta1.(2001), andKanamitsueta1.f2002).

    Inthisstudy,theNCEPDOE(DepartmentofEn

    ergy1ReanalysisIIdataset,animprovedversionof NCEPReanalysisIhavingfixederrorsandupdated

    parameterizationsofphysicalprocesses(Kanamitsuet a1.2002),wasusedfortheevaluation.NCEPIIfixed theSouthernHemispherebogusfPA0BS),whichcan contaminatesomeda_V_todayana1yses0ftheextra_

    tropica1regionsoftheSo1lthernHemisphere.The a1bed0errorovertheoceaninNCEPIisalsocor

    rectedinNCEPII.M0redetai1s0nthecorrecteder- NO.5YUETAL

    rorsandimprovementsarepresentedinKanamitsuet a1.f2002).

    TheERA40reanalysiscontainsdatafromSeptem

    ber1957toAugust2002withahorizontalresolution ofT159f125km1and60verticallevels(23stan

    dardpressurelevels1.IFSCY23r4isusedasthe atmosphericmodelforERA.4O.Acomprehensive documentationoftheECMWFforecastingsystem canbefoundontheECMWFwebsiteathttp:I1 www.ecmwf.int/research/ifsdocs/CY23r4/index.htm1. Promthebeginningoftherun.ERA40assimilated

    muchobservationaldatafromAntarcticstationsthat werenotassimilatedintoNCEPIuntiltheGlobal TelecommunicationsSystemfGTS1dataweremade available.Observationa1dataavailableforERA40

    arethosefromtheAutomaticMeteorologicalstations datafromUniversityofWisconsin(1980-1998),the AntarcticsurfacedatasetfromtheBritishAntaretic Surveyf195019991,andtheAustralianAntarcticsur

    faceandradiosondedatasetfrointheAustralianBu

    reauofMeteorologyNationaClimateCentref1947

19991(Kallbergeta1.,2004).ERA40isasecond

    generationreanalysiswithitspredecessorfERA151

    covering1979-1993fGibsoneta1.,19971.Inthispa- per.owingtotheTOVSdataintroducedintoERA4O

    in1979thevalidationwasmadefroln1979to2001. MoredescriptionsofERA40canbefoundinUppala

    eta1.f2005).

    Althoughthetwoglobalreanalysisarebasedon modelswithdifierenthorizontalandverticalresolu

    tions.theoutputsofthemodelsarearchivedona 2.5.×2.5.globallatitudeand1ongitudegridwith23 verticalpressurelevelsandatempora1resolutionof6 hrs.NCEPIIandERA40reanalysisuse3Dvaria-

    tionalassimilationf3DVAR1schemes.ForNCEPII primaryseaicedataareobtainedfromtheScanning MultichannelMicrowaveRadiometerfSMMR1and SpecialSensorMicrowave/Imager(SSM/I1:forERA

    40thesearederivedfromtheHadleyCentreGlobal SeaIceCoverandSSTdataversion1fHADISSTl1 from1957to1981andtheReynoldsoptimallyinter

    polatedseaiceconcentration(ReynoldsOI1from1982 to2002.SnowcoverdatausedforNCEPIIandERA

    40arefromtheNationalEnvironmentalSatelliteData andInformationServiceweeklyanalysesandthecli

    matologyofsnowcoverfNESDIS1andsynopticre

    portsofsnowdepth(SYNOP),respectively.These difreneesinthetworeanalysismayresultindiffer

    entoutputsoverAntarctica.

    2.2Observationaldata

    Theabovereanalysisdatasetsareevaluatedusing

    observationsfromninesurfaceweatherstationsand fourradiosondesitesatvariouslocationsinAntarc

    i153

    Fig.1.LocationsofthestationslistedinTable1 tica.Thesefbothsurfaceandupperair1dataare obtainedfromtheBritishAntarcticSurveyREADER project(http://www.antarctic.ac.uk/met/READER/) Thelocationsofthestations.togetherwiththetopog

    raphyofAntarctica.areshowninFig.1.Tablel providesgeographicinformationfortheobservational sites.aswellasinformationonthedatarecords.These stationsmostofwhichareinEastAntarcticaandall ofwhichareincoastalareas,werechosenforthestudy becauseoftheircompletenessand1ongevityofdata. Datafromthesurfacestationsarerecordedfonrtimes perdayandthelengthsofdailyrecordsvarysignifi. cantlyamongthestations,rangingfrom37yearsat Molodezna..jastationto61yearsatFaradaystation Thedataqualityisgenerallyverygood,withmissing orbadrecordsrangingfrom0.1%atBeingshausen

    to17.2%atSyowafTable11.Thesoundingdata areavailabletwiceadayandthelengthsofthedata recordsvaryamongthefourstations,fromaminimum of49years(1957-20051atSyowatoamaximumof 51years(195620061atMirny.Someofthesesound. ingdatahavebeenassimilatedintobothreanalysis datasets,sotheyarenotconsideredasindependent forcomparisonpurposes.Theresultsofcomparisons will,however,stillbevaluablebecausetheywillpro

    videabestcasescenario.

    Inthisstudy.duetothedifierenceinlengthof thetworeanalysis,andbecausethefocusofthecom

    parisonisintraseasonalandinterannualvariation,the

    dailyandmonthlymeandatafrom1979to2001are used.Thecomparisonwiththesoundingdatawas limitedtoninestandardpressurelevelsaboveground (850700,500,300200,1501OO,50and30hPa). 1154VALIDATIONOFREANALYSISDATAINANTARCTICAV0L.27 1Bellingsh62.2.S58.9.W16

    aUSen

    2Casey66.3.S110.5.E42#%

    3Davis68.6.S78.O.E%$

    4Faraday65.4.S64.4.W11%

    5Halley75.5.S26.4.W30%$

    6Mawson67.6.S62.9.E16%

    7Mirny66.5.S93.0.E30

    8

    9

    Molodez67.7.S45.9.E40

    naja

    Novolaza-70.8.S11.8.E119%

    revskaya

    10Syowa69.0.S39.6.E21%$

    1/Mar/1968

    1/Nov/2006

    56499

    31/Jan/1959

    31/Dec/2005

    68334

    10/Feb|1957

Report this document

For any questions or suggestions please email
cust-service@docsford.com