背景場誤差協(xié)方差估計(jì)技術(shù)的應(yīng)用研究.pdf

1、南京信息工程大學(xué)碩士學(xué)位論文背景場誤差協(xié)方差估計(jì)技術(shù)的應(yīng)用研究姓名：朱立娟申請學(xué)位級別：碩士專業(yè)：大氣物理與大氣環(huán)境指導(dǎo)教師：顧松山龔建東20050501AbstractThebackgrounderrorcovarianceofthetrueatmosphericisnearlyisotropic.Itisakeyproblemofvariationaldataassimilationtodescribethebackgrounder

2、orcovariance.Thebackgrounderrorcovarianceincludederrorvarianceandcharacteristicscalelengthisnotthesameindifferentareasoftheglobal.HereweestimateandtunethebackgrounderorcovariancematrixwhichadaptstoChinaregion.Theestimati

3、ngandtuningcombinesNMCmethodwhichbasesonforecastsamplesandIVmethodwhichbasesonobservationsamples.Duringthecourseofestimatingandtuningtheisotropiccovariancefunctionisadopted.Threedimensionalstructurecharactersofbackground

4、erorcovarianceofregionalnumericalmodelintheChina~isobtainedfromthestatisticresultsofNMCmethod.AndthevarianceofbackgrounderorcovarianceobtainedbyNMCmethodistunedwiththeIVmethodresults.Withuseditinthelimitedareamesoscalemo

5、delWRF3DVARthePOSTtestandsounddataassimilationvalidatesthebackgrounderorcovarianceisreasonable.OntheotherhandtheDopplerRadarVelocityisanisotropic.TheapplicationofthebackgrounderrorcovariancefunctionintheRadardataqualityc

6、ontrolispresentedwhichaimsatobtainingtheVADwindprofilefromDopplerradaraliasedvelocitydatawithoutanyextrawindinformation.ThisisbasedontheECMWFoptimalinterpolationqualitycontroltechnique.AsforthecharactersoftheDopplerRadar

7、radialvelocitydatathesinglevariablebackgrounderrorcovarianceisused.Andtheanisotropiccovariancefunctionisadoptedinthehorizontalplane.ForthetesteddataonthecircleandtiltfromwhichisretrievedtheVADwindprofilethepredigestedmul

8、tiplevariablesstatisticinterpolationtechniquejusttheCressmaninterpolationisusedtoprovideareferencevalue.AccordingtothereferencevelocitytheNyquistisdecided.Baseonthedealiasingthevelocitywecouldgetthehorizontalwindvelocity

9、fieldbytheclassicalVADtechniqueatdiferentacquiredandthenobtainverticalwindprofile.InordertoverifythemethodthecomparisonamongtheTabaryslopemethodtheBayesfunctionmethodtheYamadacontinuitymethodandthebackgrounderorcovarianc

10、efunctionmethodispresented.Theresultsshowthatthebackgrounderrorcovariancefunctionmethodisthebestone.Thecomparisonbetweenwindprofileobtainedbywindprofilerandtheoneobtainedbythebackgrounderorcovariancefunctionmetho氏andthec

11、omparisonbetweenthealiasedPPImapanddealiasedPPImaparealsopresented.Theresultshowsthatithastheabilityofobtainingaccuratewindprofilefromfolddata.Keywords:backgrounderrorcovariancecombinedNMCandIVmethod，velocityaliasingVADw