CMIP6: The Coupled Model Intercomparison Project

The Coupled Model Intercomparison Project phase 6 (CMIP6) is guided by the Grand Science Challenges of the WCRP and 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?


CMIP6 provides scientific input to the sixth assessment report of the IPCC (AR6), where MPI-M scientists act as coordinating lead authors and lead authors. CMIP6 consists a set of common experiments, the DECK (Diagnostic, Evaluation and Characterization of Klima) experiments and the Historical Simulation, common data formats and federated data storage facilities, and a set of CMIP-endorsed specific Model Intercomparison Projects (MIPs, Figure 1). Scientists of the MPI-M are involved in 15 experimental MIPs (Table 1), often taking a leading role in formulating the guiding questions and designing the experiments, for example the Paleoclimate MIP, the Radiative Forcing MIP, or the Coupled Climate-Carbon Cycle MIP. MPI-M researchers have also co-led two novel “diagnostic MIPs” (e.g., Sea Ice MIP or Dynamics and Variability MIP that focus on evaluation of specific topics and the development of novel tools.

The Max Planck Institute for Meteorology (MPI-M) participated in the recently finalized BMBF funded project DICAD that has coordinated the German contribution to the international CMIP6 archive and supported the national institutes participating in CMIP6 by developing the infrastructure for standardized data preparation and publication. DICAD partners besides MPI-M are the German Climate Computing Center (DKRZ), the German Meteorological Service (DWD), the Institute for Atmospheric Physics at the German Aerospace Center (DLR), and the Free University of Berlin.

The Max Planck Earth System Model (MPI-ESM) is the major research tool for MPI-M scientists contributing to CMIP6. Furthermore, MPI-M has recently launched its new ICON-ESM, a completely new model system that is based on advanced numerical techniques (Giorgetta et al., 2018; Korn, 2017; Modeling with ICON). ICON-ESM features unstructured grids in both ocean and atmosphere, and is well suited for modern computational infrastructure.

MPI-ESM1.2 (Mauritsen et al., 2019) is used in four configurations, which differ in resolution and the complexity of the integrated processes (Table 2). MPI-ESM-HR features higher resolution than the CMIP5 model, which has led to significant improvements in the representation of ocean and atmospheric dynamics (Müller et al., 2018). This configuration is used for historical simulations, future scenarios, and for the Decadal Climate Prediction Project (DCPP). The computationally efficient MPI-ESM-LR is used for many MIPs requiring long integrations. In “LR” the vegetation changes are calculated dynamically, which makes it suitable for MIPs with interest in land processes (e.g., LUMIP). The newly implemented nitrogen cycle was shown to reduce uncertainty in the carbon cycle feedback (Arora et al., 2020).

The High-Resolution MIP analyzes models with the highest possible resolution as part of the EU-H2020 project PRIMAVERA. Here we apply the configurations MPI-ESM-XR, with a resolution of approx. 50 km in the atmosphere and the same resolution as “HR” in the ocean, and MPI-ESM-ER with a resolution of less than 10 km in the ocean and about 100 km in the atmosphere (Gutjahr et al., 2019).

The DICAD project focused on simulations over the historical period and future projections until the end of the 21st century, for which several new shared-socioeconomic-pathway scenarios (SSP) were provided by ScenarioMIP. The SSPs take into account possible evolutions of greenhouse gas emissions and/or mitigation efforts. MPI-ESM simulates the historical evolution of global mean surface temperature in excellent agreement with the observational record (Fig. 2). This is in part due to a moderate equilibrium climate sensitivity (ECS), a feature that reflects the model’s sensitivity to increased greenhouse gas forcing. MPI-ESM’s ECS is in agreement with the most recent assessments based on observational records, but differs from several other CMIP6 models. Applying four different high-priority scenario simulations, MPI-ESM provides a range of well-separated possible futures for the end of the century. For selected scenarios ensembles of up to ten MPI-ESM simulations were carried out to include information on internal climate variability and uncertainty.

Figure 2: Time series of the global mean surface air temperature relative to the reference period 1995 to 2014 as simulated by MPI-ESM1.2 LR, HR, and ER together with the global mean surface temperature from the HadCRUT4 observational data set (yellow) for the pre-industrial control simulations (grey, black) and four CMIP6 scenarios (colors). (Figure courtesy K. Meier-Fleischer and M. Böttinger, DKRZ).

Ongoing work feeds in knowledge to AR6 by evaluating processes and feedbacks, or introducing novel methods such as unified metrics of climate sensitivity to forest cover changes, which led to new estimates global temperature change in response to deforestation (Boysen et al., 2020). While several publications document progress from CMIP5 to CMIP6, e.g. improved simulation of sea-ice loss for a given warming (SIMIP Community, 2020), or improved representation of the Quasi-Biannual Oscillation (Pohlmann et al., 2019), a manuscript with contribution from all MPI-M departments (Fiedler et al., 2020) diagnosed no general improvement in tropical precipitation across three phases of CMIP. Given the importance of tropical precipitation, the authors concluded that coarse-resolution climate models suffer from structural deficits and encourage a leap towards storm-resolving simulations.

With the newly developed ICON-ESM we have carried out CMIP6 DECK and historical experiments and publish them in the ESGF data repository. This anchors ICON-ESM in the community as a completely novel model system. A five-member ensemble of ICON-ESM historical simulations shows good agreement with the observed temperature evolution, but slightly overestimates late 20th century warming in particular in the Northern Hemisphere (Figure 3). ICON-ESM features a slightly higher climate sensitivity and its performance in comparison to MPI-ESM is presently being evaluated.

Figure 3: Time series of surface temperature over a) the globe, b) the Northern Hemisphere, and c) the Southern Hemisphere for (red-orange) the ICON-ESM historical ensemble, and observational compilations by (blue) the Goddard Institute for Space Studies Surface Temperature product, the blended Hadley Center/Climate Research Unit global temperature data set, and (blue) the NOAA NCDC historical merged land–ocean surface temperature data set The simulated global temperature is constructed using SSTs over the ocean and surface air temperatures over land.

References:

Arora, V.K., et al. (2020) Carbon-concentration and carbon-climate feedbacks in CMIP6 models, and their comparison to CMIP5 models. Biogeosciences,17, 4173–4222, https://doi.org/10.5194/bg-17-4173-2020

Boysen, L., et al. (2020) Global climate response to idealized deforestation in CMIP6 models. Under review Biogeosciences Discuss.,https://doi.org/10.5194/bg-2020-229.

Fiedler, S., et al. (2020) Simulated tropical precipitation assessed across three major phases of the Coupled Model Intercomparison Project (CMIP). Monthly Weather Review,148, 3653–3680, https://doi.org/10.1175/MWR-D-19-0404.1

Giorgetta, M. A., et al. (2018) ICON-A, the Atmosphere Component of the ICON Earth System Model: I. Model Description [Journal Article]. Journal of Advances in Modeling Earth Systems,10 (7), 1613-1637, https://doi.org/10.1029/2017ms001242.

Gutjahr, O., et al. (2019) Max Planck Institute Earth System Model (MPI-ESM1.2) for the High-Resolution Model Intercomparison Project (HighResMIP), Geosci. Model Dev.,12, 3241–3281.

Korn, P. (2017). Formulation of an unstructured grid model for global ocean dynamics. Journal of Computational Physics,339, 525-552.

Mauritsen, T., et al. (2019). Developments in the MPI-M Earth System Model version 1.2 and its response to increasing CO2. J. Adv. Model. Earth Syst.,11, 998–1038.

Mauritsen, T., and Roeckner, E. (2020) Tuning the MPI-ESM1.2 global climate model to improve the match with instrumental record warming by lowering its climate sensitivity. J. Adv. Model. Earth Syst.,12, e2019MS002037.

Müller, W.A. et al. (2018) A higher-resolution version of the Max Planck Institute Earth System Model (MPI-ESM 1.2 - HR). J. Adv. Model. Earth Syst.,10, 1383–1413.

Notz, D., SIMIP Community (2020) Arctic sea ice in CMIP6. Geophys. Res. Lett.,47, e2019GL086749.

Pohlmann, H., et al. (2019) Realistic quasi-biennial oscillation variability in historical and decadal hindcast simulations using CMIP6 forcing. Geophys. Res. Lett.,46, 14118-14125.