Yayoi Harada 1 Shinya Kobayashi 2 Yuki Kosaka 2 Jotaro Chiba 2 Takayuki Tokuhiro 2 1 Meteorological Research Institute Japan Meteorological Agency Tsukuba Japan 2 Office of Earth System Modeling Numerical Prediction Division Japan Meteorological Agency Tsukuba Japan ID: 1038625
Download Presentation The PPT/PDF document "Early results of the evaluation of the J..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1. Early results of the evaluation of the JRA-3Q reanalysisYayoi Harada1,Shinya Kobayashi2, Yuki Kosaka2, Jotaro Chiba2, Takayuki Tokuhiro21. Meteorological Research Institute / Japan Meteorological Agency, Tsukuba, Japan2. Office of Earth System Modeling / Numerical Prediction Division / Japan Meteorological Agency, Tsukuba, Japan
2. Overview of the Japanese Reanalysis for Three Quarters of a Century (JRA-3Q)
3. Japanese Reanalysis for Three Quarters of a Century (JRA-3Q)Reanalysis period: 1947 to presentSpecificationsResolution: 55 km, 60 layers (JRA-55) -> 40 km, 100 layers (JRA-3Q)Incorporating a number of improvements from the operational NWP systemOverall upgrade of physical processesNew types of observation (ground-based GNSS, hyperspectral sounders)Improved SSTCOBE-SST2 (1-degree, until around 1985) + MGDSST (0.25-degree, from around 1985 onward)Improved observational datasetsObservations newly rescued and digitised by ERA-CLIM and other projectsImproved satellite observations through reprocessingJMA’s own tropical cyclone bogus dataProduction scheduleQ3 2019: start productionQ2 2021: complete production for the 1991 – 2020 normal periodQ4 2021: Product release for the period from 1991 – presentQ1 2022: complete production for the whole period3
4. Production schedule (as of Mar 2021)4FY2015FY2016FY2017FY2018FY2019FY2020FY2021FY2022FY2023JRA-55 near-real-time productionAcquisition/QC of observations, preparation of SST and ozone data1990s–JRA-3Q near-real-time productionParallel productionProduct releaseProduction of tropical cyclone bogusConstruction of the JRA experiment/production systemPreliminary experiments (observations, boundary conditions, etc.)JRA-3Q productionWhole periodStream AUpgrade of seasonal forecast systemIntroduction of JRA-3Q climatological normalUpdate of climatologicalnormalLate 1940s–Comprehensive reportFY20141960s–1980sPreparation for connectionPreliminary reportThe Japanese financial year (FY) runs from 1 April to 31 March.Fall 2020Termination of JRA data provision from the JMA Data Dissemination SystemProvision continues from collaborative organizations.Stream AStream BStream C
5. Data assimilation system
6. Data assimilation systemJRA-55JRA-3QAnalysis periodFrom 1958 onwardFrom 1947 onwardBase systemJMA’s operational system as of December 2009JMA’s operational system as of December 2018Horizontal resolutionTL319 (~55 km)TL479 (~40 km)Vertical levels60 levels up to 0.1 hPa100 levels up to 0.01 hPaAnalysis scheme4D-Var (T106 inner resolution)4D-Var (TL319 inner resolution)Radiosonde temperature bias correctionUntil 2006: RAOBCORE V1.4From 2007 onward: RAOBCORE V1.5based on comparison with ERARISE (RICH with solar elevation dependent) v1.7.2based on comparison with surrounding stationsseasonal dependent (from 1979 onward)Satellite radiancesRTTOV-9.3 (Fast radiative transfer model developed within EUMETSAT NWP SAF)RTTOV-10.2Improved accuracyInclusion of GHGs variationsLand surface analysisOffline SiBCycle of land surface forecast from the modelSST and sea iceCOBE-SST: 1-degreeUntil 1990: COBE-SST2 (1-degree)From around 1985 onward: MGDSST (0.25-degree)Parallel production for the overlapping periodOzoneUntil 1978: ClimatologyFrom 1979 onward: MRI-CCM1 (T42L68)MRI-CCM2(TL159L64)Produced with the new model for the whole period6
7. Improvements in JMA’s operational global model since JRA-55JRA-55 (as of December 2009)Operational model (as of December 2018)ResultLongwave radiationBand-emissivity method with the diffusivity approximation2-stream absorption approximationImproved stratospheric temperature profileCloud radiationRandom overlap (shortwave)Maximum-random overlap (shortwave)Revision of optical properties of cloud water droplets Improved radiation budgetAerosolsLand and sea typessulfate, black carbon, organic carbon, sea salt and mineral dustImproved radiation budgetCumulus convectionPrognostic Arakawa-SchubertPrognostic Arakawa-SchubertImprovement of energy conservationImprovement of melting and evaporating processesImproved precipitation distributionImproved heating profileCloudSmithStratocumulus:Kawai and Inoue (2004)Smith: improvement of cloud water amount computationStratocumulus: addition of a new relative humidity threshold for occurrenceImproved cloud ice sedimentation schemeReduced dry bias in the mid troposphereSuppressed excessive stratocumuliImproved radiation budgetSurface boundary layerMonin-Obukhov similarity theoryNoniterative parameterization (Louis 1979)Monin-Obukhov similarity theoryDimensionless gradient functionsReduced excessive sensible and latent heat fluxesNon-orographic gravity wave dragRayleigh friction (above 50 hPa)Scinocca (2003)Improved representation of QBOLand surfaceSiB (Sato 1989)T: 1 (layer), moisture: 3, snow: 1Improved SiBT and moisture: 7, snow: 4 (maximum)Improved representation of surface temperature diurnal cycleSea iceSingle-layer model representing complete open water or sea ice onlyFour-layer model allowing partial sea iceReduce cold bias in the polar regions7
8. Observations
9. Observations used for JRA-3QBuilding on the JRA-55 observational datasetMaking the most of observations produced by the recent data rescue and satellite reprocessing effortsObservations at surface and upper-air stations before the year 1958 Homogenised satellite observations through reprocessingGMS/MTSAT AMVs reprocessed by JMA/MSCIncreasing satellite humidity observationsExtending the period by introducing SSM/T-2 (DMSP)Acquiring reprocessed microwave imager/humidity sounder dataAiming at reducing the dry bias in the upper/mid troposphereGenerating new tropical cyclone bogus dataUsing the JMA typhoon bogus methodSolving the problem with tropical cyclone wind retrieval (TCR) used for JRA-55Introducing new types of observing system for recent yearsGround-based GNSS (precipitable water estimation from atmospheric delay over land)Reprocessed by Dr Shoji at JMA/MRI for the period from 1994 to 2014Hyperspectral infrared sounders (high vertical resolution satellite radiances)9
10. Changes in satellite observing systems(shaded) used, (blank, pale letter) not used in JRA-5510LeoGeoInfraredMicrowaveInfrared202019701980199020002010SMSMODISMeteosatMeteosatGMSMTSATGOES WestGOES EastSSUVTPRHIRS/2SSM/ITMIAMSRSSMISAMSU-AAMSU-BMHSMSUSeaWindsASCATSSMTSCAMSNEMSSMMRSSM/T-2NSCATAMIHIRSSSHHIRS/3HIRS/4IRISSIRSIMGAIRSIASIAVHRRPMRSCRTHIRMRIRHRIRGOESSMSATSSASSRadiometerSounder(troposphere)HyperspectralSounder(Temperature)RadiometerScatterometerGNSS-ROSounder(stratosphere)Standard meridianIndian OceanWestern PacificEastern PacificAtlanticSounder(Water vapour)CrISGMISAPHIRHimawariATMSMeteosatGOES
11. Changes in satellite observing systemsUsed for JRA-3Q, (dark) addition/replacement, as of Apr 202111Reprocessed GNSS-ROLeoGeoInfraredMicrowaveInfrared202019701980199020002010SMSMODISMeteosatGMSSSUVTPRHIRS/2AMSU-ASCAMSNEMSHIRSSSHHIRS/3HIRS/4IRISSIRSIMGAVHRRPMRSCRTHIRMRIRHRIRGOESSMSATSSASSRadiometerSounder(troposphere)HyperspectralSounder(Temperature)RadiometerScatterometerGNSS-ROSounder(stratosphere)Standard meridianIndian OceanWestern PacificEastern PacificAtlanticSounder(Water vapour)ATMSMeteosatGOES WestGOES EastHimawariSSM/T-2SMMRTMIAMSRSSMISSeaWindsASCATNSCATAMSU-BMHSAIRSIASICrISReprocessed ocean surface windsreprocessed AMVreprocessed AMVrecalibrationSSMTMSUrecalibrationrecalibrationSAPHIRAMIGMISSM/IMeteosatGOESMTSATMWRIWindSat
12. Early results of the evaluation of JRA-3Q
13. Departures of radiosonde temperatures13JRA-3Q Obs – BG JRA-55 Obs – BG JRA-25 Obs – BG Lower stratosphereJump after Mt Pinatubo eruption in 1991Weaker stratospheric warming in JRA-3QUpper troposphereGlobal mean (left) drawing near zeroDiminished warm bias in JRA-3QLower troposphereGlobal mean (left) reducedMitigated cold bias in JRA-3QGlobal mean (K)RMS(K)
14. RMS errors of 5-day forecasts14Geopotential height at 500 hPaWind vector in the TropicsThe thinning interval for TOVS was shortened (250 km -> 125 km) and the BG error covariances (except for log(q)) were inflated by 11 % in Oct 1990.The impact of these tunings has been fully reflected on the scores since Oct 1991 because each value represents the average for the last 12 months.The scores under shading will be updated once short-range forecasts are completed for the Stream B period (up to Dec 1990).
15. Precipitation distribution (1991-2000 years mean)15JRA-3Q (comparison with JRA-55, JRA-25):Reduced overestimation in the tropics, particularly in the ITCZDiminished underestimation over AmazonGPCPJRA-3Q – GPCPJRA-55 – GPCPJRA-25 – GPCPMERRA – GPCPMERRA2 – GPCPCFSR – GPCPERA5 – GPCP
16. Precipitation distribution (2001-2010 years mean)16JRA-3Q (comparison with JRA-55, JRA-25):Reduced overestimation in the tropics, particularly in the ITCZDiminished underestimation over AmazonGPCPJRA-3Q – GPCPJRA-55 – GPCPJRA-25 – GPCPMERRA – GPCPMERRA2 – GPCPCFSR – GPCPERA5 – GPCP
17. Precipitation distribution (2011-2015 years mean)17JRA-3Q (comparison with JRA-55, JRA-25):Reduced overestimation in the tropics, particularly in the ITCZDiminished underestimation over AmazonGPCPJRA-3Q – GPCPJRA-55 – GPCPJRA-25 – GPCPMERRA – GPCPMERRA2 – GPCPERA5 – GPCP
18. Long-term variability of precipitation in the tropics18Precipitation (22S-22N mean, mm/day) 12-month running mean precipitation and spatial anomaly correlation coefficients against GPCP V2.3. Anomalies for each dataset were defined relative to their own climatological monthly means over 1991–2010 (1991-2001 for ERA-40, 1997-2010 for TRMM).Precipitation anomaly from 1991-2010 mean Spatial anomaly correlations vs. GPCP V2.3JRA-3Q (comparison with JRA-55, JRA-25):Reduced overestimation and relatively stable inter-annual variabilityHigher spatial anomaly correlation throughout the period JRA-3QJRA-55JRA-3QJRA-3QJRA-55
19. Specific humidity in the troposphere (1991-2015 years mean)19JRA-3Q (comparison with JRA-55):Reduced dry bias seen in the mid-troposphere of JRA-55, particularly in the equatorial regionJRA-3QJRA-3Q – JRA-55JRA-3Q – ERA5Specific humidity @ 600 hPa (5S-5N mean, 10-3 kg/kg) JRA-3Q – MERRA2JRA-3QJRA-55
20. Diabatic heating rate (2015-2018 JJA mean)JRA-3QConvectiveheating rate [K day-1]Large scale condensation heating rate [K day-1]JRA-55JRA-3QJRA-55JRA-3Q (comparison with JRA-55):Reduced top-heavy convective heating rate. Improved of Q1-QR vertical profile. Vertical profile of Q1-QR (0-15N, 60-150E mean) during JJA 1998 – 2007. Q1-QR for reanalyses are calculated as the sum of convective, large scale condensation, vertical diffusion heating rates. JRA-3Q – JRA-55JRA-3Q – JRA-55
21. Diabatic heating rate (2015-2018 DJF mean)JRA-3QConvectiveheating rate [K day-1]Large scale condensation heating rate [K day-1]JRA-55JRA-3QJRA-55JRA-3Q (comparison with JRA-55):Reduced top-heavy convective heating rate. Improved of Q1-QR vertical profile. Vertical profile of Q1-QR (0-15N, 60-150E mean) during JJA 1998 – 2007. Q1-QR for reanalyses are calculated as the sum of convective, large scale condensation, vertical diffusion heating rates. JRA-3Q – JRA-55JRA-3Q – JRA-55
22. Latent heat flux (1991-2015 years mean)22JRA-3Q (comparison with JRA-55, JRA-25):Reduced overestimation in the tropics Relatively large positive biases still remain in the tropical South Pacific Ocean.OAFluxJRA-3Q – OAFluxJRA-55 – OAFluxJRA-25 – OAFluxMERRA – OAFluxMERRA2 – OAFluxERA5 – OAFlux
23. Sensible heat flux (1991-2015 years mean)23JRA-3Q (comparison with JRA-55, JRA-25):Reduced overestimation except for the west coast of the continentsOAFluxJRA-3Q – OAFluxJRA-55 – OAFluxJRA-25 – OAFluxMERRA – OAFluxMERRA2 – OAFluxERA5 – OAFlux
24. Upward long-wave radiation at TOA (2002-2015 years mean)24JRA-3Q (comparison with JRA-55, JRA-25):Reduced positive bias around the maritime continent, central Africa and AmazonOverall weak positive biases remain. CERES-EBAFJRA-3Q – CERESJRA-55 – CERESJRA-25 – CERESMERRA – CERESMERRA2 – CERESERA5 – CERES
25. Specific humidity in the middle atmosphere (2005-2018DJF mean) 25JRA-3Q (comparison with JRA-55):Reduced overestimation of water vapor around tropopause and in the stratosphere due to the implementation of methane oxidation scheme based on Untch and Simmons (1999)Weak underestimations remain in the mesosphere of the summer Hemisphere MLS/AuraJRA-3Q – MLS/AuraJRA-55 – MLS/Aura
26. Time variation of specific humidity in the equatorial (5S-5N) middle atmosphere26MLS/AuraJRA-3Q – MLS/AuraJRA-55 – MLS/AuraJRA-3Q (comparison with JRA-55): Reduced overestimation of water vapor around 10 hPaThe representation of tape recorder signal in the lower stratosphere seems to be insufficient. JRA-3QJRA-55
27. Time variation of specific humidity in the polar (75-90N) middle atmosphere27MLS/AuraJRA-3Q – MLS/AuraJRA-55 – MLS/AuraJRA-3Q (comparison with JRA-55): Reduced overestimation of water vapor in the lower stratosphereImproved representation of annual cycle both in the upper stratosphere and the mesosphere JRA-3QJRA-55
28. Ozone mixing ratio in the middle atmosphere (2005-2018 DJF mean)28JRA-3Q (comparison with JRA-55):Reduced underestimation in the stratosphere, particularly in the tropics due to the usage of the upgraded chemistry-climate model (MRI-CCM2, Deushi and Shibata 2010)Weak underestimations are found in the polar region of the winter hemispheric mesosphere MLS/AuraJRA-3Q – MLS/AuraJRA-55 – MLS/Aura
29. Time variation of ozone mixing ratio in the equatorial (5S-5N) middle atmosphere29JRA-3Q (comparison with JRA-55):Reduced underestimation in the stratosphereMLS/AuraJRA-3Q – MLS/AuraJRA-55 – MLS/AuraJRA-3QJRA-55
30. Time variation of ozone mixing ratio in the polar (75-90N) middle atmosphere30JRA-3Q (comparison with JRA-55):Reduced underestimation in the upper stratosphereWeak underestimations are seen in the winter hemispheric mesosphere MLS/AuraJRA-3Q – MLS/AuraJRA-55 – MLS/AuraJRA-3QJRA-55
31. Official product
32. Outline of official productFile format: GRIB2Total volume: approximately 350 TB (1947-2020)Horizontal resolution: approximately 40 km (TL479 model grid)Temporal resolution32CategoryModel grid analysis1.25-deg analysisModel grid forecast1.25-deg forecastSurface (two-dimensional instantaneous) fields6 (6)6 (6)1 (3)1 (3)Two-dimensional average diagnostic fields1 (3)1 (3)Two-dimensional extreme fields1 (3)Land surface instantaneous fields6 (6)6 (6)1 (3)1 (3)Land surface average diagnostic fields1 (3)1 (3)Snow depth analysis fields24 (24)24 (24)Model level instantaneous fields6 (6)3 (6)Model level average diagnostic fields6 (6)Isobaric instantaneous fields6 (-)6 (6)3 (-)3 (6)Isobaric average diagnostic fields6 (6)Isentropic fields6 (6)6 (6)Unit is hour. Values in parentheses indicate the temporal resolution for JRA-55.
33. Official product (vertical coordinates)33Model level60 layers -> 100 layersRaise of the model top0.1 hPa -> 0.01 hPaIncrease of vertical resolutionIsobaric level37 layers -> 45 layersAddition around the tropical tropopause85, 60, 40 hPaAddition to the mesosphere(based on the user survey conducted by S-RIP)Isentropic level21 layers -> 36 layersAddition to the stratosphere and mesosphere(based on the user survey conducted by S-RIP)S-RIP: SPARC Reanalysis Intercomparison Project