prsentation Bilan Prospective GAME

1
Plans from Météo-France / CNRM and CERFACS for
CMIP5
D. Salas-Mélia, S. Planton, A.-L. Gibelin and A.
Voldoire (CNRM) L. Terray and C. Cassou (CERFACS)
2
CNRM-CM3
Introduction
Ozone chemistry MOBIDIC
Aerosol DataGHG
Zonal Statistics
O3
10 years
Atmosphere ARPEGE-Climat, T42, L45
Continental Surfaces ISBA (cycle C)
24h
24h
24h
OASIS
Sea Ice
Ocean OPA 2, L31
Rivers TRIP
24h
3
CMIP3 / IPCC AR4
Simulated SST (global mean, year average)
4
ENSEMBLES stream 2 simulations

• Recent developments since CMIP3
• Include ARPEGE-Climat 4 (new dynamical core)
• Sulphate aerosols first indirect effect
• Improve ARPEGE-OPA coupling (energy conservation)
• Changing land-use new vegetation map
• New ozone linear scheme
• Impact of volcanic eruptions (stratospheric
  aerosols)

5
CNRM-3.3 (2008)

• Last millenium climate simulation (CERFACS)
• Simulations of 1860-2000 with CNRM-CM3.3 (DMI,
  Denmark), ENSEMBLES project

6
ENSEMBLES Stream 2 Control exp. / 20C / 21C-A1B
Simulated SST (global mean, year average)
7
1. Which model ?

• AOGCM two versions may be available higher
  priority
• CM4LR Lower resolution ARPEGE-Climat 5
  (2.8-L31) NEMO (2-L31) our current model
• CM4HR Higher resolution ARPEGE-Climat 5
  (1.4-L31) NEMO (1-L31) (promising trials with
  ARPEGE 4, 1-L31 / NEMO 0.5-L31)
• ESM uncertain
• PMIP tier 1, possibly part of tier 2
• CFMIP
• Off-line COSP calculations, on-line if time
  allows

8
2. Which experiments ?

• Short term decadal
• Run by CERFACS, initialization anomalies,
  10-member ensembles
• Higher resolution ARPEGE-Climat 5 (1.4-L31) NEMO
  (1-L31)
• Initial ocean conditions provided by ECMWF
  (obtained from NEMO 1 runs)
• Long term simulations
• Run by CNRM
• At least all  mandatory  most  very high
  priority runs  will be done

9
3. Interactive chemistry, coupled chemistry ?

• Interactive ozone (linear scheme)
• As for CMIP3, but formulation of the scheme is
  improved
• Aerosols will be prescribed from RCPs

10
4. When are we starting ?

• Simulations will start in early 2010
• Computing system will be upgraded in March 2009
  (NEC)
• Current version of CNRM-CM a 1000-yr simulation
  takes 4 months to complete on current NEC machine

11
5. Constraints

• People
• Technical supports for running simulations ok
• Lack of physics developers
• Data storage (publicly available data through
  DODS server)
• A question how shall we transfer the data to
  PCMDI ?
• Exchanging 1Tb disks with PCMDI ?