Periodic Reporting for period 2 - EERIE (European Eddy-RIch ESMs)
Reporting period: 2024-05-01 to 2025-09-30
EERIE (European Eddy-Rich Earth System Models) addresses a fundamental gap in climate science: while the real ocean is rich in mesoscale eddies, classical CMIP-class climate models do not resolve them. This limits our ability to simulate climate variability, extremes and potential tipping behaviour in a warming world. EERIE develops and exploits a new generation of global, coupled eddy-rich Earth System Models that explicitly represent ocean mesoscale dynamics, combining kilometre- to eddy-scale simulations with novel experimental designs, advanced HPC optimisation, FAIR data infrastructures and machine-learning methods. First key results demonstrate the feasibility and scientific value of this approach. The Phase-1 simulations have been successfully completed, delivering the first coordinated, multi-model set of century-scale eddy-rich global climate simulations. A novel data-management and access framework enables efficient handling, analysis and reuse of these unprecedented datasets. Initial evaluation shows that ocean eddies are realistically represented across models, both in open-ocean and ice-covered regions, providing a robust foundation for improved climate projections, attribution of extremes, and downstream use in European climate services and Digital Twins of the Earth.
During the first and second reporting periods, EERIE has delivered major advances in the development, execution and exploitation of global eddy-rich Earth System Models (ESMs), with a focus on the climatic role of ocean mesoscale processes.
A central achievement of RP2 has been the successful production of the first coordinated, multi-model set of century-scale, fully coupled eddy-rich climate simulations, including kilometre-scale atmosphere (~9 km) and eddy-rich ocean resolutions. These simulations were completed across several modelling systems and now form the backbone of scientific analysis in WPs 5–12.
To enable efficient exploitation of these unprecedented datasets, EERIE established a modern, scalable and FAIR data ecosystem. Phase-1 outputs were quality controlled, standardised and published through ESGF, CEDA and WDCC, with additional fast access via eerie.cloud using Zarr, STAC catalogues and virtual datasets. Integrated visualisation and analysis tools substantially reduce time-to-science and support ML workflows.
Scientific evaluation during RP2 shows that EERIE models realistically capture key characteristics of the ocean mesoscale while identifying systematic weaknesses in regions with complex dynamics. These insights directly informed a refined Phase-2 strategy, shifting towards small ensembles of eddy-rich simulations to robustly quantify uncertainty and internal variability.
Further achievements include demonstrable improvements in large-scale circulation, air–sea coupling and the simulation of high-impact events, significant gains in computational efficiency, and the delivery of the first stable hybrid coupled climate model combining a dynamical ocean with an ML-based atmosphere.
A central achievement of RP2 has been the successful production of the first coordinated, multi-model set of century-scale, fully coupled eddy-rich climate simulations, including kilometre-scale atmosphere (~9 km) and eddy-rich ocean resolutions. These simulations were completed across several modelling systems and now form the backbone of scientific analysis in WPs 5–12.
To enable efficient exploitation of these unprecedented datasets, EERIE established a modern, scalable and FAIR data ecosystem. Phase-1 outputs were quality controlled, standardised and published through ESGF, CEDA and WDCC, with additional fast access via eerie.cloud using Zarr, STAC catalogues and virtual datasets. Integrated visualisation and analysis tools substantially reduce time-to-science and support ML workflows.
Scientific evaluation during RP2 shows that EERIE models realistically capture key characteristics of the ocean mesoscale while identifying systematic weaknesses in regions with complex dynamics. These insights directly informed a refined Phase-2 strategy, shifting towards small ensembles of eddy-rich simulations to robustly quantify uncertainty and internal variability.
Further achievements include demonstrable improvements in large-scale circulation, air–sea coupling and the simulation of high-impact events, significant gains in computational efficiency, and the delivery of the first stable hybrid coupled climate model combining a dynamical ocean with an ML-based atmosphere.
EERIE has delivered results that go decisively beyond the current state of the art in global climate modelling.
For the first time, multi-decadal to century-scale, fully coupled eddy-rich climate simulations are available across multiple independent ESMs. Earlier studies were limited to short integrations, single models or uncoupled configurations. EERIE therefore enables robust assessment of how ocean mesoscale processes influence climate variability, extremes, circulation changes and potential tipping behaviour.
The project demonstrates that explicitly resolving mesoscale dynamics leads to systematic improvements in large-scale circulation features, coastal sea-surface temperatures, water-mass pathways and air–sea interaction processes that directly affect regional climate and extremes. EERIE further shows that mesoscale impacts on the atmosphere arise primarily through indirect pathways linked to heat and freshwater transports, refining long-standing hypotheses with quantitative evidence.
By resolving mesoscale-driven variability, EERIE improves estimates of background natural variability, a prerequisite for reliable regional climate attribution and risk assessment. This represents a substantial advance over conventional CMIP-class models.
Methodologically, EERIE introduces novel experimental designs extending CMIP capabilities, including ensemble-based eddy-rich strategies and targeted sensitivity experiments. Through leadership of the HighResMIP2 design for CMIP7, EERIE is shaping next-generation global high-resolution intercomparisons. The integration of eddy-rich simulations with ML-based emulators and hybrid models opens new pathways to deliver high-fidelity climate information at manageable computational cost.
For the first time, multi-decadal to century-scale, fully coupled eddy-rich climate simulations are available across multiple independent ESMs. Earlier studies were limited to short integrations, single models or uncoupled configurations. EERIE therefore enables robust assessment of how ocean mesoscale processes influence climate variability, extremes, circulation changes and potential tipping behaviour.
The project demonstrates that explicitly resolving mesoscale dynamics leads to systematic improvements in large-scale circulation features, coastal sea-surface temperatures, water-mass pathways and air–sea interaction processes that directly affect regional climate and extremes. EERIE further shows that mesoscale impacts on the atmosphere arise primarily through indirect pathways linked to heat and freshwater transports, refining long-standing hypotheses with quantitative evidence.
By resolving mesoscale-driven variability, EERIE improves estimates of background natural variability, a prerequisite for reliable regional climate attribution and risk assessment. This represents a substantial advance over conventional CMIP-class models.
Methodologically, EERIE introduces novel experimental designs extending CMIP capabilities, including ensemble-based eddy-rich strategies and targeted sensitivity experiments. Through leadership of the HighResMIP2 design for CMIP7, EERIE is shaping next-generation global high-resolution intercomparisons. The integration of eddy-rich simulations with ML-based emulators and hybrid models opens new pathways to deliver high-fidelity climate information at manageable computational cost.