MiKlip

MiKlip

MiKlip (German: Mittelfristige Klima-Prognosen) was a German research program with the aim to develop a decadal climate prediction system (Marotzke et al., 2016). The project ended in 2020. Towards the end of the program, the global forecasting system has been transferred to DWD and it will be used for the operational production of decadal predictions. MiKlip was organized into 5 modules:





The global prediction system uses the Max-Planck Institute for Meteorology Earth-System Model MPI-ESM.

Within Module C the regionalization of the global predictions was performed with COSMO-CLM. During the first phase of MiKlip two additional RCMs (REMO and WRF) were also tested.

To estimate the skill of the predictions, large sets of initialized decadal hindcasts are performed, usually covering the period after 1960 with annually starting 10-year simulations. In contrast to the un-initialized simulations (historical simulations in CMIP) the hindcast derive their initial conditions from observations via assimilation experiments using atmospheric and oceanic re-analysis.

Four generations of hindcasts have been generated in MiKlip up to now:

1. baseline0 (= CMIP5)

  • GCM:  MPI-ESM-LR – Atmosphere: ECHAM6 T63L47; Ocean: MPI-OM  1.5° L4

  • Initialization -  Ocean: Anomalies of temperature and salinity from an MPI-OM forced by NCEP re-analysis

  • Ensemble-size - 10 member every 5-years (1960, 1965,...), 3 members in-between years (1961, 1962,...)

  • Starting years: 1960 – 2011 (e.g. simulation period decadal1960 = 01.01.1961 – 31.12.1970) ·

2. baseline1

  • GCM:  MPI-ESM-LR and MPI-ESM-LR (Atmosphere: T63L95, Ocean: 0.4° TP L40)

  • Initialization – Ocean – anomalies from ORA S4 ocean re-analysis, Atmosphere – full field ERA

  • Ensemble size – 10 members (LR), 5 members (MR)

  • Starting years: 1960 – 2016 (e.g. simulation period decadal1960 = 01.01.1961 – 31.12.1970)

3. prototype

  • GCM:  MPI-ESM-LR

  • Initialization – Ocean – full field from ORA S4 and GECCO2 ocean re-analysis, Atmosphere – full field ERA

  • Ensemble size – 15 (ORA S4 init.) + 15 (GECCO2 init.)

  • Starting years: 1960 – 2014 (e.g. simulation period decadal1960 = 01.01.1961 – 31.12.1970)

4. preop (pre-operational)

  • GCM:  MPI-ESM-HR  (Atm: T127L95, Ocean: 0.4° TP L40)

  • Initialization – as in baseline1

  • Ensemble size – 5 members (CMIP5 forcings), 10 members (CMIP6 forcings)

  • Starting dates: 01.11.1960 – 01.11.2017



Additionally, experiments have been performed to investigate the climate variability and trend over the whole 20th century for Europe, including re-analysis driven simulations based on NCEP 20CR and ERA20CC, MPI_ESM-LR decadal hindcasts, and historical simulations.

Model domain

For the downscaling for Europe the EURO-CORDEX domain is used:

For Europe, the simulations are performed on a rotated grid with the pole at 162°W longitude and 39.25°N latitude. The domain covers approximately the region from about 22°W to 45°E longitude and 27°N to 72°N latitude. The resolutions applied are 0.44° (EUR-44) and 0.22° (EUR-22) and, for Central Europe, 0.0625° (CEUR-0625).

The recommended COSMO-CLM setup is used for the simulations:

  • COSMO4.8_clm19 (CCLM4-8) for baseline0 and baseline1

  • COSMO5.0_clm9 (CCLM5-0) for prototype and preop

Simulations performed with COSMO-CLM

Decadal Hindcasts

Ens. Generation

Model version

Grid

Starting years

Ensemble members

# Simulation years

Ens. Generation

Model version

Grid

Starting years

Ensemble members

# Simulation years

baseline0

COSMO-CLM4-8

EUR-22

5 (1960, 1970, 1980 1990 2000)

10

500



COSMO-CLM4-8

CEUR-0625

5 (1960, 1970, 1980 1990 2000)

4

200

baseline1

COSMO-CLM4-8

EUR-22

5 (1960, 1970, 1980 1990 2000)

10

500



COSMO-CLM4-8

EUR-44

57 (annual 1960 – 2016)

10

5700

prototype                     

COSMO-CLM5-0

EUR-22

5 (1960, 1970, 1980 1990 2000)

1

50

preop

COSMO-CLM5-0

EUR-22

57 (58) (annual 1960 – 2017)

5 (10)

2850 (5800)

20C_EUR

COSMO-CLM5-0

EUR-22

100 (annual 1910 – 2009)

3                    

3000

Evaluation Simulations

Ens. Generation

Model version

Grid

Period

Ensemble members

# Simulation years

Ens. Generation

Model version

Grid

Period

Ensemble members

# Simulation years

ERA40                         

COSMO-CLM4-8

EUR-22      

1961 - 2002                            

1

42

ERAInterim

COSMO-CLM4-8

EUR-22

1979 - 2010              

1

32

ERA40+ERAInterim

COSMO-CLM4-8

EUR-22

1961 - 2010

1

49

ERA40+ERAInterim

COSMO-CLM4-8

EUR-44

1961 - 2010

1

49

ERA40+ERAInterim

COSMO-CLM5-0

EUR-22

1961 - 2016

1

56

NCEP 20CR

COSMO-CLM5-0

EUR-22   

1900 - 2009

3

110

ERA20CF

COSMO-CLM5-0

EUR-22

1900 - 2010

1

111

Un-initialized (historical) simulations

GCM

Model version

Grid

Period

Ensemble members

# Simulation years

GCM

Model version

Grid

Period

Ensemble members

# Simulation years

MPI-ESM-LR (CMIP5)        

COSMO-CLM4-8

EUR-22

1961 - 2010

1

49

MPI-ESM-LR (CMIP5)

COSMO-CLM4-8

EUR-44

1961 - 2010

7

49

MPI-ESM-HR (CMIP5)                    

COSMO-CLM5-0

EUR-22

1900 – 2030 (1961-2030)

rcp45 after 2005

3 (2)

131 (70)

Additional Experiments

Name

Project

Details

Name

Project

Details

Africa

MiKlip I DEPARTURE

Decadal predictions of the West-African Monsoon and tropical Atlantic cyclogenesis

3 RCMs (COSMO-CLM4-8, REMO, WRF)

Resolution 0.44° (extended CORDEX Africa

4 Decades, 3 Members

Central America/North-Atlantic (CANA)

MiKlip I MesoTel

2-way nested coupling of MPI-ESM and COSMO-CLM

Regionally coupled ocean

MiKlip II C1-WP1

Interactive coupling of COSMO-CLM5-0/NEMO for Mediterranean- Nordic, and Baltic-Sea

Alternative SVAT

MiKlip II C1-WP2

Decadal hindcasts (preop generation) with COSMO-CLM5-0-Weg3D

Many more experiments have been performed including statistical dynamical downscaling with COSMO-CLM.

Contributing Partners MiKlip Phase II

  • Karlsruhe Institute of Technology: Christoph Kottmeier, Hendrik Feldmann, Hans-Jürgen Panitz, Gerd Schädler, Joaquim G. Pinto, Marcus Breil, Natalie Laube, Julia Mömken, Benjamin Buldmann, Fanni Kelemen

  • Deutscher Wetterdienst: Barbara Früh, Sascha Brand

  • University Frankfurt: Bodo Ahrens, Anika Obermann, Nora Leps, Naveed Akhtar, Cristina Primo Ramos

  • University Cologne: Mark Reyers

  • University Würzburg: Heiko Paeth, Jingmin LI

(Support)Support