Over its full duration, nextGEMS has successfully created, tested, and demonstrated two storm-resolving Earth system models that mark a fundamental step forward in climate simulation. One is based on the ICON modelling framework, the other on the Integrated Forecasting System (IFS) of ECMWF. Both were extended and optimized to exploit the full capacity of Europe’s high-performance computing infrastructure, becoming the first Earth system models to run efficiently on Europe’s most performant and energy-efficient supercomputers (LUMI and JEDI). New, hierarchical, cache-optimized output handling methods were developed to manage and distribute data efficiently and enable a global user community to access unprecedented data volumes. Through its pioneering use of Hackathons nextGEMS has demonstrated the efficiency of remote data handling, advanced best practices, and built an integrated international community of users outside of the major modelling centers.
The development of the HAM lite aerosol model allowed aerosol forcing estimates to be performed at a horizontal grid spacing of 5 km and enabled a more physical treatment of mineral dust emissions. The resulting global aerosol forcing (–0.4 W / m2) which is on the lower end of established assessments. ICON was also coupled to the carbon cycle and ocean biogeochemistry, enabling the first five-year global simulations of coupled carbon–climate feedbacks. A new urban scheme was introduced in IFS-FESOM and compared favourably to observed urban heat-island effects – something only possible with models capable of representing the scale of the urban environment. Dedicated experiments explored cloud feedbacks and climate sensitivity under +4 K and 4× CO2 forcing across a range of horizontal grid spacings, from 20 km to 2.5 km. Positive cloud feedbacks and climate sensitivity increased with finer resolution pointing approaching the upper envelope of estimates from more conventional models. For the first time the effect of organization on the response of clouds to warming has been studied, and the SR-ESMs could be shown to converge toward the observations at grid spacings of 2.5 km. Simulations were performed and analysed to explore air-sea interaction in the tropics in association with tropical instability waves, tropical storms, ENSO, and the mean tropical climate. A strong signature of the ocean on the atmosphere is found for tropical instability waves. For ENSO variability an asymmetry arises to differences in the depth of the mixed layer, making warm phases with a shallow mixed layer more sensitive to surface fluxes and precipitation variability that could be better simulated at km scales. The project demonstrated that SR-ESMs improve the representation of blocking, together with more realistic jets and storm tracks. Biases in block size and persistence are reduced and improved blocking frequency patterns over the North Atlantic and North Pacific are evident.
nextGEMS' key results were the ESM Development, the simulation outputs, and publications and methods. All were exploited in different ways depending on the target groups. The developed models are now used beyond the project in EERIE, WarmWorld, and DestinE. Workflows and HPC strategies such as chunking, HEALPix, and intake catalogs support Digital-Twin development and wider community access. The simulation outputs are available via WDCC and Intake catalogs for all interested communities. Results are widely disseminated through peer-reviewed papers, videos, and policy briefs.