The EU project European Eddy RIch Earth System Models (EERIE) aims to advance kilometer-scale Earth System Models (ESMs) to reduce biases associated with low-resolution climate simulations. Its goal is to develop centennial-scale ESMs that explicitly resolve ocean mesoscale processes, thereby improving the representation of long-term climate evolution, variability, extremes, and potential tipping points. One of these models, IFS-FESOM2-SR, couples the ECMWFs Integrated Forecast System (IFS) atmosphere (9 km resolution) with the FESOM2.5 ocean model (minimum 5 km resolution). The FESOM2.5 ocean model employs an NG5 unstructured triangular grid with 70 depth levels, achieving ~5 km resolution in eddy-rich mid- and high-latitudes and ~13 km in the tropics (Rackow et al., 2025). Its sea-ice component is FESIM (Danilov et al., 2015). The atmospheric model, IFS cycle 48r1 from ECMWF, uses a Tco1279 (~10 km) octahedral grid with 137 vertical levels. The setup follows Rackow et al. (2025) except for deep convection, where the operational IFS scheme is used instead of the modified reduced cloud-base mass flux version. Following the HighResMIP protocol (Haarsma et al., 2016), the main simulations were preceded by a 50-year spin-up period using 1950 CMIP6 forcing. From the spin-up’s final state, two simulations were launched in parallel: a control run and a historical run using CMIP6 forcings. According to the HighResMIP protocol, the control simulation aimed to assess any potential drift within the simulation, enabling us to exclude the influence of such drift in order to better understand the impact of changes in radiative forcing over time. After completion of the historical simulation, the experiment was extended along the SSP2-4.5 scenario pathway until 2050 to estimate near-future climate change using a long-term climate simulation at the kilometer scale. For this purpose, CMIP6 scenario forcings were used. Tropospheric aerosol estimates are based on the MACv2 aerosol forcing, which was subsequently adjusted to ensure compatibility with the CONFESS aerosol forcing used during the historical simulation.

The EU project European Eddy RIch Earth System Models (EERIE) aims to advance kilometer-scale Earth System Models (ESMs) to reduce biases associated with low-resolution climate simulations. Its goal is to develop centennial-scale ESMs that explicitly resolve ocean mesoscale processes, thereby improving the representation of long-term climate evolution, variability, extremes, and potential tipping points. One of these models, IFS-FESOM2-SR, couples the ECMWFs Integrated Forecast System (IFS) atmosphere (9 km resolution) with the FESOM2.5 ocean model (minimum 5 km resolution). The FESOM2.5 ocean model employs an NG5 unstructured triangular grid with 70 depth levels, achieving ~5 km resolution in eddy-rich mid- and high-latitudes and ~13 km in the tropics (Rackow et al., 2025). Its sea-ice component is FESIM (Danilov et al., 2015). The atmospheric model, IFS cycle 48r1 from ECMWF, uses a Tco1279 (~10 km) octahedral grid with 137 vertical levels. The setup follows Rackow et al. (2025) except for deep convection, where the operational IFS scheme is used instead of the modified reduced cloud-base mass flux version. Following the HighResMIP protocol (Haarsma et al., 2016), the main simulations were preceded by a 50-year spin-up period using 1950 CMIP6 forcing. From the spin-up’s final state, two simulations were launched in parallel: a control run and a historical run using CMIP6 forcings. According to the HighResMIP protocol, the control simulation aimed to assess any potential drift within the simulation, enabling us to exclude the influence of such drift in order to better understand the impact of changes in radiative forcing over time. The historical simulation employed CMIP6 historical forcing spanning from 1950 to 2014, running for a total of 65 years. In the context of tropospheric aerosols, we employed the CONFESS aerosol forcing, which is available from 1970 onwards and is applied in five-year epochs. For the period preceding 1970 (from 1950 to 1969), we generated epochs by replicating the 1970 aerosol forcing. This generation of aerosol forcing was conducted as part of the EU project DestinE (Destination Earth), from which additional model developments and adaptations are being integrated. These efforts aim to prepare the IFS-FESOM for conducting multidecadal climate simulations.

The EU project European Eddy-Rich Earth System Models (EERIE) aims to advance kilometre-scale Earth System Models (ESMs) to reduce biases associated with low-resolution climate simulations. Its overarching goal is to develop centennial-scale ESMs that explicitly resolve ocean mesoscale processes, thereby improving the representation of long-term climate evolution, variability, extremes, and potential tipping points. One of these novel models, the IFS–NEMO–ER coupled system, combines version v4.0.7 of the ocean model NEMO (Madec et al., 2019) and the sea-ice model SI3 (Vancoppenolle et al., 2023) with an upgraded version of the atmospheric model IFS cycle 48r1 (i.e. DE_CY48R1.0_EERIE_20240726). Two configurations are available, with intermediate and high horizontal resolutions. The high-resolution configuration employs a triangular–cubic–octahedral (Tco1279) atmospheric grid (~9–10 km) with 137 vertical levels and a tripolar global orthogonal curvilinear ocean grid (eORCA12, 1/12°) with 75 depth levels. The intermediate-resolution configuration uses a Tco399 atmospheric grid (~28 km, 137 vertical levels) and the eORCA025 ocean grid (1/4°, 75 depth levels). The intermediate-resolution setup is primarily used for model development and tuning and serves as a lower-resolution counterpart for EERIE experiments. All model components are coupled using a single-executable approach (Mogensen et al., 2012), enabling highly efficient, low-latency data exchange. Coupling is performed at an hourly frequency, with the exchange of key variables such as sea surface temperature, surface fluxes, sea-ice cover, and ocean currents. The atmosphere–sea-ice coupling follows a new thermodynamic strategy in which SI3 provides sea-ice concentration, albedo, and ice surface temperature directly to IFS. To equilibrate the coupled system before conducting the production experiments, a 60-year spin-up simulation was first performed under fixed 1950 radiative forcing. The final state of this spin-up was used to provide dynamically balanced initial conditions for the subsequent control-1950 and hist-1950 simulations. The spin-up was initialized with ocean and sea-ice states from EN4 observations representative of 1950 (Good et al., 2013), while atmospheric initial conditions were taken from ERA5 in 2020. Although this atmospheric state is warmer than that of 1950, its impact on the coupled system is negligible due to the atmosphere’s low heat capacity. The historical simulations use CMIP6 historical forcing for the period 1950–2014, covering a total length of 65 years. Tropospheric and volcanic aerosol forcing is prescribed using the CONFESS aerosol forcing, which is available from 1970 onward and provided in five-year epochs. For the earlier period (1950–1969), aerosol forcing is constructed by replicating the 1970 aerosol forcing. The aerosol forcing was generated within the EU Destination Earth (DestinE) project, from which further model developments and adaptations are currently being incorporated. These efforts are intended to prepare the IFS–NEMO system for multidecadal climate simulations.

The EU project European Eddy RIch Earth System Models (EERIE) is developing a new generation of Earth System Models (ESMs) that explicitly resolve ocean mesoscale dynamics, an essential but still poorly explored part of the climate system. By using recent advances in computing and model design, EERIE aims to improve long-term climate simulations, including variability, extremes, and potential tipping points influenced by mesoscale ocean processes. ICON in Sapphire configuration is one of these new models. Developed at the Max Planck Institute for Meteorology, ICON couples the atmosphere, land, ocean, and sea ice at kilometer-scale resolution. It resolves deep atmospheric convection and captures mesoscale to sub-mesoscale ocean eddies, with the option to refine the global ocean grid locally as a “computational telescope.” The atmospheric component uses a nonhydrostatic icosahedral C grid with a hybrid sigma-z vertical coordinate and parameterizes only unresolved processes (radiation, microphysics, turbulence). The ocean component shares the same grid and solves the hydrostatic Boussinesq equations, using only a subset of parameterizations such as vertical mixing and velocity dissipation. Sea ice is included via FESIM dynamics and a simplified thermodynamic scheme. Ocean biogeochemistry is represented by HAMOCC6, simulating more than 20 tracers. The land component, JSBACH 4, provides surface fluxes and simplified hydrology with prescribed vegetation. All components are coupled through the YAC coupler (v2.4.2). The main simulations were preceded by a 40-year spin-up period using 1950 CMIP6 forcing. From the spin-up’s final state, two parallel simulations were started: a 100-year control run and a historical run. The control run is used to identify and quantify model drift, ensuring that any long-term changes in the historical simulation could be attributed to variations in radiative forcing rather than internal drift. The historical simulation employed CMIP6 historical forcing spanning from 1950 to 2014, running for a total of 65 years. Only volcanic aerosol forcing was taken from CMIP5.

The EU project European Eddy RIch Earth System Models (EERIE) is developing a new generation of Earth System Models (ESMs) that explicitly resolve ocean mesoscale dynamics, an essential but still poorly explored part of the climate system. By using recent advances in computing and model design, EERIE aims to improve long-term climate simulations, including variability, extremes, and potential tipping points influenced by mesoscale ocean processes. ICON in Sapphire configuration is one of these new models. Developed at the Max Planck Institute for Meteorology, ICON couples the atmosphere, land, ocean, and sea ice at kilometer-scale resolution. It resolves deep atmospheric convection and captures mesoscale to sub-mesoscale ocean eddies, with the option to refine the global ocean grid locally as a “computational telescope.” The atmospheric component uses a nonhydrostatic icosahedral C grid with a hybrid sigma-z vertical coordinate and parameterizes only unresolved processes (radiation, microphysics, turbulence). The ocean component shares the same grid and solves the hydrostatic Boussinesq equations, using only a subset of parameterizations such as vertical mixing and velocity dissipation. Sea ice is included via FESIM dynamics and a simplified thermodynamic scheme. Ocean biogeochemistry is represented by HAMOCC6, simulating more than 20 tracers. The land component, JSBACH 4, provides surface fluxes and simplified hydrology with prescribed vegetation. All components are coupled through the YAC coupler (v2.4.2). The main simulations were preceded by a 40-year spin-up period using 1950 CMIP6 forcing. From the spin-up’s final state, two parallel simulations were started: a 100-year control run and a historical run. The control run is used to identify and quantify model drift, ensuring that any long-term changes in the historical simulation could be attributed to variations in radiative forcing rather than internal drift.

The EU project European Eddy RIch Earth System Models (EERIE) is developing a new generation of Earth System Models (ESMs) that explicitly resolve ocean mesoscale dynamics, an essential but still poorly explored part of the climate system. By using recent advances in computing and model design, EERIE aims to improve long-term climate simulations, including variability, extremes, and potential tipping points influenced by mesoscale ocean processes. ICON in Sapphire configuration is one of these new models. Developed at the Max Planck Institute for Meteorology, ICON couples the atmosphere, land, ocean, and sea ice at kilometer-scale resolution. It resolves deep atmospheric convection and captures mesoscale to sub-mesoscale ocean eddies, with the option to refine the global ocean grid locally as a “computational telescope.” The atmospheric component uses a nonhydrostatic icosahedral C grid with a hybrid sigma-z vertical coordinate and parameterizes only unresolved processes (radiation, microphysics, turbulence). The ocean component shares the same grid and solves the hydrostatic Boussinesq equations, using only a subset of parameterizations such as vertical mixing and velocity dissipation. Sea ice is included via FESIM dynamics and a simplified thermodynamic scheme. Ocean biogeochemistry is represented by HAMOCC6, simulating more than 20 tracers. The land component, JSBACH 4, provides surface fluxes and simplified hydrology with prescribed vegetation. All components are coupled through the YAC coupler (v2.4.2). The main simulations were preceded by a 40-year spin-up period using 1950 CMIP6 forcing. From the spin-up’s final state, two parallel simulations were started: a 100-year control run and a historical run. The control run is used to identify and quantify model drift, ensuring that any long-term changes in the historical simulation could be attributed to variations in radiative forcing rather than internal drift. After completing the historical simulation, the experiment was extended along the SSP2-4.5 pathway to 2050, using CMIP6 scenario forcings. This extension enables estimates of near-future climate change from a long-term, kilometer-scale simulation.