ECMWF數(shù)值預(yù)報(bào)模式簡(jiǎn)介精品課件氣象預(yù)報(bào)的四個(gè)方面問(wèn)題當(dāng)前的天氣或氣候信息—完備的綜合觀測(cè)系統(tǒng)

完善的觀測(cè)系統(tǒng)；資料信息識(shí)別；資料的綜合處理（同化）天氣或氣候的演變規(guī)律—從資料得到新認(rèn)識(shí)

從資料得到新認(rèn)識(shí)；反映大氣運(yùn)動(dòng)數(shù)學(xué)物理規(guī)律的微分方程組。外力和強(qiáng)迫的變化—地形和邊界強(qiáng)迫；太陽(yáng)常數(shù)；引力等從已知預(yù)報(bào)未來(lái)的手段—完備的數(shù)值模式；承載數(shù)值模式計(jì)算、顯示和通訊平臺(tái)駕馭觀測(cè)、資料分析和數(shù)值模式發(fā)展和應(yīng)用的人才隊(duì)伍精品課件大氣圈水圈冰晶圈/冰凍圈生物圈巖石圈人類(lèi)圈精品課件紛繁多樣數(shù)值預(yù)報(bào)產(chǎn)品

-什么會(huì)有差別？精品課件分辨率不同模式的表述的物理過(guò)程有差異

計(jì)算方案、網(wǎng)格、變量分布等再分析資料分析的初值能代表大氣實(shí)況嗎？

精品課件OrographyandResolutionsGSMT213(60km)RSM(20km)MSM(10km)Orographiceffectsarebettercapturedbyhigherresolutionmodels.ThesurfaceparameterssuchasTsurfmightbepredictedmorerealisticallybythosemodels.精品課件ECMWF致力于2016-2025數(shù)值預(yù)報(bào)發(fā)展戰(zhàn)略：

利用集合預(yù)報(bào)方法提前2周預(yù)報(bào)高影響天氣事件

提前4周，無(wú)縫隙地預(yù)報(bào)大尺度系統(tǒng)形勢(shì)和系統(tǒng)移動(dòng)

提前1年預(yù)測(cè)全球范圍的異常狀況

精品課件

研究和更豐富的知識(shí)基于集合預(yù)報(bào)的分析和預(yù)報(bào)技術(shù)如何實(shí)現(xiàn)目標(biāo)：ObservationsHighresolutionensemble

Earth-system

ScalabilityFunding

People

（引自ErlandK?llén，“Earthsystemmodellingforseamlessprediction”ECMWFAnnualSeminar2016）精品課件FutureEarthSystemmodelandassimilationdevelopments

ErlandK?llén,ECMWF

精品課件精品課件Oceanmodel(NEMO)Wavemodel(ECWAM)Atmospheric

model(IFS)AirdensityGustinessNeutralwindRoughnessCurrentsSeasurfacetemperatureAllconfigurationsEnsembleFCSolarandnonsolarfluxes,E-PTowardsafullycoupledsystem(currentlyonlyoperationalinEPS)ORCA1_Z42TCo1279/TCo63914km/28km9km/18kmEnsemblesystemsonly:-Mediumrangeforecast-Monthlyforecast-SeasonalforecastEveryIFStimestepEverycouplingtimestep(1or3hours)Singleexecutable精品課件Oceanmodel(NEMO)Wavemodel(ECWAM)Atmospheric

model(IFS)AirdensityGustinessNeutralwindRoughnessTurbulentenergyStokesdriftCurrentsSeasurfacetemperatureAllconfigurationsStressEnsembleFCSolarandnonsolarfluxes,E-PTowardsacoupledsystemORCA1_Z42TCo1279/TCo63914km/28km9km/18kmSingleexecutableOperationalfromday0since2013精品課件Oceanmodel(NEMO)IceconcentrationWavemodel(ECWAM)Atmospheric

model(IFS)AirdensityGustinessNeutralwindRoughnessTurbulentenergyStokesdriftCurrentsIcemodel(LIM)SeasurfacetemperatureIceconcentrationAllconfigurationsStressEnsembleFCfutureoperationalSolarandnonsolarfluxes,E-PTowardsacoupledsystemORCA0.25°_Z75AddingactiveseaicemodelImplementation:End2016.Singleexecutable精品課件CallingsequenceofthesingleexecutableSimplifiedflowchartofthecoupledmodel,heretwotimestepsareshown.Inreality,IFS/WAMcouplingeveryIFStimestep,butCalltoNEMOeveryhour(or3hours)withaveragedaccumulatedfluxes.14精品課件ImpactofResolutionontropicalcycloneforecastForinstanceTyphoonHaiyan:forecastsfrom4th,5thand6thNovember2013,0UTCallfromoperationalanalysis.Black:estimatedfromobservationsRed:oldoperationalEnsembleresolution(32km)Blue:oldoperationalHRESconfiguration(16km)Green:experimental:newHRESconfiguration(10km)minMSLP(hPa)900950TyphoonHaiyanatpeakintensityonNovember7,2013精品課件ImpactofCouplingontropicalcycloneforecastForinstanceTyphoonNeoguri:forecastsfrom6July2014,0UTCBlack:estimatedfromobservationsGreen:oldoperationalHRESconfiguration(uncoupled)(16km)Red:experimental:16kmcoupledtoNEMO(ORCA025_Z75)Blue:16kmcoupledtoNEMO+newphysicsminMSLP(hPa)900950NeoguriaffectingOkinawaonJuly8,2014精品課件ClimatereanalysesforthecoupledearthmodelECMWFcoupledEarthmodelformedium-rangeweatherforecastingNewcoupledassimilationsystem(CERA)forthecoupledEarthmodel:atmosphericandoceanobservationsassimilatedsimultaneouslyoceanobservationscanimpactatmosphericestimateandconverselyCERA-20Creanalysisinproduction(1900-2010)KarlandTrenberth2003landwavesseaiceatmosphereoceansECMWFcoupledEarthmodel精品課件Coupledassimilationsystem(CERA)EDAvariationalapproachwitha24-hourwindowthatassimilatessimultaneouslyatmosphericandoceanobservationscoupledmodelcomputesobservationmisfitsineachouteriterationatmosphericandoceanincrementsarecomputedinparalleltocorrecttheinitialstateSSTcomputedinNEMOandconstrainedbyrelaxationanalysisdynamicallyconsistentwithrespecttothecoupledmodel精品課件Conclusions:ECMWFhasacoupledatmosphere-wave-oceancirculationforecastingsystem,currentlyoperationalintheEnsemblePredictionSystem.Workisongoingonusingahigherresolutionoceancomponents(ORCA025z75)plannedforendof2016intheEnsembleforecastsandlaterintheHighresolutionsystem.Thereisaclearbenefitincouplingthedifferentmodels,butitcreatesnewchallengesindeterminingwhatphysicalparametersneedtobeexchanged.Furthermore,modelparameterisationsmightneedrevisiting.精品課件EC模式動(dòng)力框架垂直采用η坐標(biāo)，U=ucosθ

；V=vcosθ精品課件精品課件精品課件對(duì)比球坐標(biāo)系中的基本方程組曲率項(xiàng)精品課件重力不是指向地心，等位勢(shì)面也不是球面。精品課件非靜力學(xué)模式靜力學(xué)模式靜力學(xué)平衡：運(yùn)動(dòng)的垂直尺度遠(yuǎn)小于水平尺度的情況下成立模式的分類(lèi)精品課件基于連續(xù)方程的模式分類(lèi)非靜力學(xué)模式根據(jù)模式中是否包含聲波，分為：滯（非）彈性（anelastic)模式

假定大氣不可壓，濾掉聲波彈性(elastic)模式

大氣可壓、由散度預(yù)報(bào)氣壓的變化，但聲波需要特殊的處理

精品課件根據(jù)連續(xù)方程的近似程度，滯彈性和彈性模式又可進(jìn)行不同的分類(lèi)：滯彈性模式不可壓縮模式滯彈性模式彈性模式準(zhǔn)可壓模式完全可壓模式精品課件數(shù)值計(jì)算方案

大氣運(yùn)動(dòng)方程組是一套復(fù)雜的非線(xiàn)性方程組，目前，還沒(méi)有理論解存在，只能借助數(shù)值方法求解。數(shù)值方法有很多，目前氣象上用的主要用的是差分方法、譜方法、有限元法等。差分方法：就是在離散的網(wǎng)格點(diǎn)上，以差商代替微商，求解微分方程的方法。精品課件關(guān)于有限差分與譜方法的圖例有限差分譜方法精品課件精品課件

|||

n-1nn+1從物理上考慮，差分的精度還與采樣點(diǎn)密—即網(wǎng)格，與要素在采樣區(qū)的變化分布有關(guān)。即采樣數(shù)據(jù)是否能有代表性。中央差精確值采樣值混淆誤差精品課件離散網(wǎng)格所能表示的最短波長(zhǎng)上表中當(dāng)L/x=2時(shí)，即兩倍格距波，R=0，誤差100