nextGEMS: Wrapping Up Four Years at the Frontiers of Climate Modelling

When nextGEMS began in 2021, it set out with an ambitious goal: to bring global, kilometre-scale Earth system modelling from dream to reality. Four years later, as the project closes at the end of August 2025, the results speak for themselves — and for the hundreds of scientists, engineers, and collaborators who made it possible.

A Vision Realised

For the first time, multi-decadal global simulations at the kilometre scale — coupling atmosphere, ocean, and land — have been run with two different models: ICON and IFS-FESOM. Both now achieve an energetically consistent climate, a challenge that has plagued even long-established climate models, and can simulate decades in a matter of weeks, thanks to throughputs of up to 600 simulated days per day.

These breakthroughs lay the groundwork for the Climate Change Adaptation Digital Twin in the European Commission’s Destination Earth initiative and set a benchmark for future climate research.

The Power of Community

From the start, nextGEMS set out to break down silos between climate science, high-performance computing, and model development. Over 25 institutions from Africa, Asia, Europe, and North America joined forces, meeting online and — critically — in person at six hackathons.

Participant numbers grew from around 80 in the first cycle to over 130 in the third and fourth, with each meeting strengthening inter-institutional cooperation and cross-disciplinary exchange. The project’s open-data ethos, supported by tools like easy.gems, unified Intake catalogues, and example notebooks, made it easier than ever for participants to analyse terabytes of simulation output directly on the Levante supercomputer.

Hackathons: Where the Magic Happened

The six hackathons — in Berlin, Vienna, Madrid, Hamburg, Wageningen, and Stockholm — were much more than coding sprints. They were week-long collaborative laboratories where model developers, climate scientists, HPC specialists, and sectoral users sat side by side.

Working directly on Levante meant no need to transfer enormous datasets; ideas could be tested in real time. Feedback from these sessions fed directly into the next model cycle, making the hackathons the heartbeat of nextGEMS.

Nurturing the Next Generation

Early-career scientists (ECSs) played a central role—not only as observers, but also as coders, analysts, and co-authors. For many, nextGEMS offered their first opportunity to contribute to a major international modelling initiative.

Hackathons, in particular, created a unique environment where ECSs could learn directly from senior scientists. They received guidance on project design and research methods that helped them avoid wasted effort while fostering creativity and innovation. These events also provided valuable opportunities to build personal networks. Meeting hundreds of international researchers laid the foundation for long-term collaborations and partnerships, while also helping young scientists advance their ongoing projects. One hackathon participant, for example, used the event to distribute surveys and recruit interview partners for research on the human factor in developing the nextGEMS Earth system model.

Beyond hackathons, nextGEMS enabled ECSs to strengthen both scientific and technical skills. These included recognizing the limitations of models in representing the climate system, managing complex data, and analyzing results from multiple high-resolution global simulations. As one early-career scientist from the University of Hamburg reflected:

“Engaging with a large community and diverse model applications encouraged me to think creatively and broaden my perspective. These experiences were essential in shaping my understanding of the field.”

By providing space for learning, mentoring, and collaboration, nextGEMS has ensured that its legacy lies not only in the data and models it produced but also in the new generation of scientists it has prepared to lead the field forward.

Science Meets Society

However, nextGEMS didn’t only speak to scientists. Industry stakeholders from energy providers and fisheries joined hackathons to explore how kilometre-scale simulations could inform their planning. For example:

• Renewable energy operators examined high-resolution wind and solar data for better resource forecasting.
• Fisheries managers explored ocean currents and temperature changes that affect ecosystems.

At the 2025 Final Countdown Hackathon in Stockholm, a participant from the renewable energy company Statkraft described the experience as especially valuable. The informal setting allowed direct conversations with climate scientists and other participants about the challenges encountered when working with climate data and their opinions on its quality and usability. Additionally, the close collaboration with nextGEMS’ technical experts offered guidance on navigating specific technical aspects of the datasets, such as zoom levels and the Xarray grid.

Technical Milestones

Among the most notable achievements:

• Multi-decadal, energetically consistent km-scale simulations for 2020–2050 under the SSP3-7.0 scenario.
• Throughput leaps — from 17 simulated days per day in early ICON runs to over 400, and from 40 to 600 in IFS-FESOM.
• HEALPix output grids for direct model-to-model comparison.
• Harmonised data formats and common variable sets, enabling faster, more transparent analysis.

Scientific Insights

Working at kilometre scale allowed nextGEMS to explore climate processes in unprecedented detail:

• Tropical rainbelt dynamics and their connection to sea-surface temperature patterns, as studied by Segura et al., 2024, Bao et al., 2024, and Hohenegger et al., 2023.
• Soil moisture–precipitation feedbacks, which are key for drought prediction and were investigated by Lee & Hohenegger, 2024 and Hohenegger et al., 2023, among others.
• Stratocumulus cloud fields play a crucial role in the Earth’s radiation balance and are discussed in publications, such as those by Nowak et al., 2025 and Waclawczyk et al., 2022.

These insights are not only scientifically exciting — they have direct relevance for climate-sensitive sectors and adaptation strategies worldwide.

Looking Ahead

While the Horizon 2020 funding chapter closes, nextGEMS’ legacy is just beginning. The data, tools, and community it built will feed into Destination Earth’s digital twins, inform policy-relevant research, and continue to train the next generation of climate modellers.

The km-scale era has arrived — and the people who made nextGEMS possible have ensured it’s here to stay.

For those who want to dive deeper, we invite you to visit our multimedia library, where you can watch videos exploring the project’s key outcomes in more detail — from behind-the-scenes insights at our hackathons to explanations of what kilometre-scale modelling means for science and society. You can also explore the full breadth of our scientific output in our Zenodo community. Among these, we especially recommend the newly released paper by Segura et al. (2025), “nextGEMS: entering the era of kilometre-scale Earth system modelling”, which captures the technical breakthroughs, collaborative spirit, and scientific discoveries that define the project.