Regional ReAnalyses Overview, background and needs Differerent types of reanalyses Multi models – Ensembles Uncertainties – estimated and real ? Use of reanalyses and uncertainties User interaction Copernicus Climate Change Services (C3S)
2
3 All observations 1961-2015 (c.o. Klaus Zimmerman)
Gridding of observations Statistical interpolation of observations only 2 m temperature and precipitation and more CRU data sets global ~ 50 km 1901 - E-OBS data sets Europe land ~25 km 1955 - GPCC global precipitation ~ 50 km 110 years Several regional-national data sets Alpine gridded data set precipitation 5 km Nordic gridded data set 1 km 1981- UK, KNMI, SMHI, MF, DWD …….. 4 FE12 Seminar, 12.08.2015 L. Bach
Post−1950 data Obs. per day 1 TT RH SD/FS/RR 2 Infilling >1950 3 DP SLP 4 WS/WD ³ 5 50 Length (years) <10 and <1950 10−19 20−29 30−39 40 40−49 complemented by French ³ 50 and Swedish 60's data 30 And Norwegian,Swedish and Catalonian data 20 −10 0 10 20 30 40 5
Data assimilation of Observations 6
7 MET Office hybrid 4D-VAR/3D-VAR ensembles
Background From FGGE first GARP global experiment Additional radiosondes and drifting buoys – 1978/79 level IIIb ”reanalysis” Bengtsson and Shukla (1988) Observing system performance, predictability, general circulation, low frequency variability ERA-15 global reanalysis to cover this 8 FE12 Seminar, 12.08.2015 L. Bach
Global reanalyses – ERA-15 – Observing systems developments ERA-40, Interim, ERA-20C. and ERA-5 Copernicus Helsinki WMO conference – need for regional reanalyses North American efforts Japanese NCEP Climate reanalysis JRA-55 NCEP 20CR NASA MERRA-2 NCEP RRA ASR 2000-10 EU Commission and SPACE programme DAMOCLES Arctic ocean and ”lightly” coupled atmosphere EURO4M UERRA COSMO RRA 6 km (1995-2015) MÉRA 2.5 km (1981-2015 ) CMA Spain, Italy, India Netherlands Australia and more 9 FE12 Seminar, 12.08.2015 L. Bach
MÉRA: ALADIN-HIRLAM NWP System • AHNS used by 26 NHMS Europe & North Africa • Used operationally by Met Éireann since 2011 • HARMONIE-AROME configuration (cycle 38h1.2) – 2.5km horizontal grid-spacing – 65 levels with highest level at 10hPa – ALADIN non-hydrostatic dynamics – MESO-NH mesoscale physics – SURFEX surface model – 3D-Var upper-air data assimilation – OI surface data assimilation Met Éireann
Forecast performance PMSL Temperature Corrected Temperature Wind Met Éireann
Climate change service EU Regulation (911/2010) specifies that “access to information for climate change monitoring in support of mitigation and adaptation policies” shall be included in the GMES service component. A Climate Change service must meet the Global Climate Observing System goals: Monitor the climate system Detect and attribute climate change Assess impacts of, and support adaptation to, climate variability & change from Strahlendorff (2012) from Strahlendorff (2012)
Climate change service consultation process: expert group (2010-2011), ‘Helsinki GMES climate’ conference (Jul 2011), GMES user forum (Nov 2011). 6 th FP7 space call (closed since 21 Nov. 2012) EC funding: 26 M€ re-analysis (global and EU regional) quality assurance for ECVs climate indicator toolbox Attribution products data access
One of 5 pre-operational Copernicus Projects ERA-CLIM2 European Reanalysis of the Global Climate System UERRA Uncertainties in Ensembles of Regional ReAnalyses QA4ECV Quality Assurance for Essential Climate Variables CLIPC A Climate Information Portal for Copernicus EUCLEIA European Climate and weather events: interpretation and attribution 14 FE12 Seminar, 12.08.2015 L. Bach
Copernicus Regional Reanalysis for Europe UERRA and C3S 322 Lot1 (PRECISE) Climate Change For the Arctjc: C3S 322 Lot2 (CARRA) UERRA : Grant Agreement 607193 EU FP7 SPACE 2013- 1
1 6 Data available at: https://software.ecmwf.int/wiki/display/UER
1 7 HARMONIE (HIRLAM ALADIN Regional/Mesoscale Operational NWP In Europe) Forecasts – 30 hour forecasts at 00 and 12 – 1 hour resolution up to 6 hours, 3 hour resolution up to 30 hours Run at ECMWF – Several streams (5-10 years) with 4 months overlap – ~180 milj. SBUs – ~1200 Tb data of which ~350 Tb is stored in MARS 5 years with HARMONIE-ALARO – 2006-2010 – Preparation for the longer re-analysis – Mini ensemble for uncertainty estimates
1 8 Verification of T2m, winter 1962 1972 1982 Red: UERRA Green: ERA 1992 2002 2012
MESAN downscaling surface/2D analyses MESAN – Swedish surface analysis system combined and run at 11 km – First guess from forecast model and special first guess for precipitation – Non-isotropic structure functions SMHI HIRLAM 11 km AROME HARMONIE 2.5 19
MESCAN downscaling ensemble analyses MESAN – CANARI – Swedish and French/ALADIN surface analysis system combined and run at 5.5 km SMHI HARMONIE ALADIN @ 11 km – ALADIN physics -> – ALARO physics -> • Interpolate first guess 6 hour forecast • Interpolate analysis → run ALADIN 5.5 km forecast – SURFEX soil scheme – ISBA soil scheme » Perturb observations » Reduced network 20
21 Curtesy Francesco Isotta (Meteo-Swiss)
Spread of 2m temperature for March 1979 from the Met Office system Ensemble 4D-VAR reanalysis 22
Rank (Talagrand) Histogram DWD MET Office CERA-20C 23
Ensemble spread No overlap in spread • Bias too large…? • …spread too small? Actually pretty good! From Besselaar and van den Schrier (2017)
Lessons learned from the pre- operational Copernicus projects Need to make potential users aware or the possibilities of modern reanalysis data More work to assess and reduce long term Biases Need to develop data assimilation and model based reanalyis systems different from operational systems (especially regional) since they are of different resolutions and domains 25 FE12 Seminar, 12.08.2015 L. Bach
Lessons learned from the pre- operational Copernicus projects Consistent and combined information (atmosphere, cryosphere, atmospheric chemistry) needed Need for improved satellite data Critical in-situ data need to be rescued, quality controlled and made available Intensify communication and engagement with society 26 FE12 Seminar, 12.08.2015 L. Bach
Final points UERRA HARMONIE (SMHI) and MESCAN (MF) lead to C3S_322 lot 1 European Regional operational Reanalysis Also an Arctic Regional Reanalysis How to balance : – Resolution – domain (EU, region) – ensemble Uncertainties still considerable and difficult Need to improve on top of operational systems and input observations rescued and Qced! www.uerra.eu 27
Recommend
More recommend
