The coupled model intercomparison project phase 6 (CMIP6) is an international endeavour to better understand past, present and future climate changes. Referring to the Grand Science Challenges of the World Climate Research Programme (WCRP), CMIP6 aims to answer three broad questions:

(i) How does the Earth system respond to forcing?,

(ii) What are the origins and consequences of systematic model biases?, and

(iii) How can we assess future climate changes given climate variability

, predictability and uncertainties in scenarios?


To tackle these questions, CMIP6 consists of three major elements: First, a handful of common experiments, the DECK (Diagnostic, Evaluation and Characterization of Klima experiments) and the CMIP Historical Simulation (1850 - near-present). Second, common data formats, a federated data storage and documentation to facilitate the distribution of model outputs and the evaluation of the multi model ensemble. Third, a complementary set of CMIP-Endorsed Model Intercomparison Projects (MIPs) aiming at specific questions.


The Max Planck Institute for Meteorology (MPI-M) will participate in CMIP6 with different configurations of the Max Planck Earth System model (MPI-ESM), based on the known ECHAM6/MPIOM (MPI-ESM-1) circulation models as well as on the new ICON models (ICON-ESM). All models will run the DECK experiments (AMIP simulation, pre-industrial control simulation, abrupt 4x CO2 simulation, and 1%/year CO2 increase simulation) as well as the historical simulation, which serve as an “entry card” to CMIP6.

* The MPI-ESM1.2-XR is part of HighResMIP and will not perform the full DECK.

MIPs and the MPI-M

The MPI-M is involved in 17 MIPs, where additional experiments are conducted to investigate the variability and predictability of the Earth system, to better understand systematic model biases, and to see how the climate reacts to forcing, in particular to future anthropogenic forcing which is investigated by performing scenario simulations. In many of these MIPs, the scientists of the MPI-M take a leading role in in formulating the guiding questions and setting up the experimental design, such as the Paleoclimate MIP (PMIP), the Radiative Forcing MIP (RFMIP) or the Couple Climate-Carbon Cycle MIP (C4MIP).

Of the 17 MIPs, two are classified as "diagnostic MIPs" (DynVarMIP and SIMIP) in which no specific experiments are proposed, but additional output from the DECK and selected MIP experiment is requested.

The scope, scientific questions, experimental design and requested output data of the MIPs are described in detail in a special issue of Geoscientific Model Development (GMD):

Model Development towards CMIP6 

In comparison to CMIP5, the MPI-ESM-MR has been replaced with the MPI-ESM1.2-HR which offers a higher resolution in the atmosphere component (from T63 to T127). Furthermore, all model components have been improved. This includes the implementation of the new PSrad radiation scheme and the removal of energy leakages in ECHAM6.3 (Pincus and Stevens, 2013). YASSO was introduced as the new soil carbon model and a coupled nitrogen cycle has been implemented in JSBACH (both relevant only for MPI-ESM1.2-LR where dynamic vegetation is enabled).

Major technical work went into further optimization of the code on the DKRZ MISTRAL machine as well as the implementation of new CMIP6 diagnostics to fulfill the CMIP6 output requirements. Moreover, the recommended input datasets — such as land-use change, ozone, solar, and stratospheric aerosol forcing — for CMIP6 had to be processed in order to be used with the MPI-ESM and ICON-ESM. In particular, the new MACv2-SP scheme for tropospheric aerosol forcing was developed at the MPI-M and is now a standard tool for CMIP6 Models (Stevens et al., 2017).

The ICON-ESM is a completely new Earth System Model which was not available for CMIP5. The ICON-ESM will perform the DECK and selected simulations in different MIPs. This provides the unique opportunity for the ICON-ESM to be evaluated against international state-of-the-art coupled climate models.

CMIP6 Simulations with the MPI-ESM

The development of the MPI-ESM1.2 is completed and apart from possible small bugfixes, no major change of the code is planned for CMIP6. Both model configurations have been tuned to match our best understanding of the 20th century climate. Part of the DECK simulations (500 years of the pre-industrial control simulations and the 1%/year-CO2 increase simulation) as well as 5 members of the historical simulations have been finished with the MPI-ESM1.2-HR in autumn 2017. 











In terms of global mean surface temperature anomalies, the MPI-ESM1.2-HR with the new CMIP6 forcing data is in good agreement with observations (left figure). In comparison to the CMIP5 version of the MPI-ESM, the MPI-ESM1.2-HR shows a reduced, and hence more realistic, cooling following large volcanic eruptions as well as a better representation of the mid-20th century warming. The Atlantic Meridional Overturning Circulation shows a realistic magnitude of about 16 Sv, comparable to the CMIP5 MPI-ESM-MR, which has the same resolution in the ocean component MPIOM as the MPI-ESM1.2-HR.

In their evaluation of the MPI-ESM1.2-HR, Müller et al. (2017) diagnose improvements in terms of bias reduction and representation of dynamical features — such as storm tracks and blocking situation in the North Atlantic sector — in comparison with the MPI-ESM1.2-LR. The benefit of finer model resolution is explored further in HighResMIP using an even higher resolution configuration (MPI-ESM1.2-XR).


Special Issue of Geoscientific Model Development (GMD) containing the general CMIP6 overview as well as papers of all MIPs and forcing datasets:

Müller, W. A., et al (2018), A high resolution version of the  Max Planck Institute Earth System Model (MPI-ESM1.2-HR), J. Adv. Model. Earth Syst.,

Pincus, R. and B. Stevens (2013), Path to accuracy for radiation parameterizations in atmospheric models, J. Adv. Model. Earth Syst., 5(2), 225-233, doi:10.1002/jame.20027

Stevens, B., et al (2017), MACv2-SP: A parameterization of anthropogenic aerosol optical properties and an associated Twomey effect for use in CMIP6, Geosci. Model Dev., 10(1), 433-452, doi:10.5194/gmd-10-433-2017