HClimRep

HClimRep researchers contribute to MLESM 2026 in Bonn

Agnes Lohs — Thu, 08 Oct 2026 13:49:33 +0000

Members of the HClimRep project joined researchers from around the world in Bonn for MLESM 2026, a two-part event featuring a hackathon and a workshop in August. More than 100 researchers from the Earth system science and machine learning (ML) communities came together to explore how machine learning can help address emerging challenges in Earth system research.

The event was hosted by the Institute of Computer Science at the University of Bonn and supported by the Center for Earth System Observation and Computational Analysis (CESOC) and the HClimRep project.

The event consisted of two parts: a hackathon from 19 to 21 August, followed by a workshop from 24 to 26 August. Held for the second consecutive year, the hackathon brought participants together for three days of coding, experimentation, and collaborative work on concrete challenges at the intersection of machine learning, Earth system science, and high-performance computing. Members of HClimRep played an active role in organising the event and contributed as lecturers and tutors during the hackathon.

A challenge for the hackathon: Integrating SST into weather and climate models

One of the five hackathon challenges was organised and tutored by researchers from the HClimRep project and members of the ESDE team from Forschungszentrum Jülich. The challenge, “Integrating SST into weather and climate models,” explored how sea surface temperature (SST) can be incorporated into a modern, data-driven weather model to better capture long-term warming trends.

Working with the WeatherGenerator model, the teams compared two different integration strategies and developed their own evaluation routines to assess their performance. The challenge was supervised by Savvas Melidonis, Jifeng Wang, Jehangir Awan, and Ankit Patnala. The model’s response to the integrated information was measured and successfully analysed.

Input from the experts

The hackathon was accompanied by a series of lectures, including a contribution from Martin Schultz, lead PI of HClimRep, on the fundamentals of conventional and machine-learning-based weather modelling. This and other sessions provided participants with additional background and context for their three days of hands-on coding and data analysis.

Connecting research through weather and climate

Following the hackathon, the fourth MLESM workshop provided a platform for scientists to present their latest research, exchange ideas, and discuss the challenges and opportunities of machine learning for Earth system modelling. Topics covered a broad range of areas, including machine-learning-based numerical weather prediction, high-resolution forecasting, climate models, Earth system components, observations, foundation models, and the evaluation of AI-based models.

Savvas Melidonis, Head of the HClimRep project, presented a poster entitled HClimRep, a foundation model for capturing atmosphere, ocean, and sea-ice interactions. The poster highlighted the project’s work on developing the first foundation model for climate science that simultaneously integrates atmospheric, oceanic, and sea-ice data and observations.

The HClimRep consortium also contributed an oral presentation on HClimRep-Ocean (OceanRep), delivered by Kacper Nowak, a PhD student at AWI.

Each time, the MLESM event reflects the growing interest in machine learning as a tool for weather and climate research, while also highlighting the importance of interdisciplinary collaboration between the Earth system science, mathematics, and computational science communities. For researchers from HClimRep, this year’s event provided another opportunity to contribute to these discussions, share their work, and exchange perspectives on the development of AI-based approaches for climate modelling.

Der Beitrag HClimRep researchers contribute to MLESM 2026 in Bonn erschien zuerst auf HClimRep.

HClimRep at the 9th ENES HPC Workshop

Agnes Lohs — Mon, 15 Jun 2026 12:07:57 +0000

HClimRep was represented by Project Coordinator Savvas Melidonis at the 9th ENES HPC Workshop on “HPC for High-Resolution Climate and Weather Modelling”, held from 19–21 May 2026 at the Swedish Meteorological and Hydrological Institute (SMHI) in Norrköping, Sweden. The workshop brought together leading researchers, software developers, HPC specialists, and industry experts to discuss the growing importance of future weather and climate data, the challenges of high-performance computing in a rapidly changing technological landscape, and the emerging role of artificial intelligence in Earth system modelling.

Organised by ESiWACE3, the ENES HPC Workshop is a key forum within the European climate and weather modelling community. This year’s programme focused on collaboration across modelling centres and HPC facilities, advances in performance and accelerator technologies, AI and machine learning applications, and the management and use of increasingly large climate datasets. As Earth system models continue to move towards higher resolutions and exascale computing, these discussions are becoming increasingly important for the future of climate research and operational forecasting.

HClimRep’s participation provided an opportunity to engage with experts at the forefront of climate and weather computing, exchange knowledge on emerging technologies, and strengthen connections with the wider European modelling community. Such engagement is essential for ensuring that HClimRep remains aligned with the latest developments in high-performance computing and data-intensive climate science, supporting the project’s mission to improve the accessibility, usability, and long-term value of high-resolution climate information.

Der Beitrag HClimRep at the 9th ENES HPC Workshop erschien zuerst auf HClimRep.

HClimRep annual in-person meeting 2026, Hamburg

Agnes Lohs — Wed, 25 Mar 2026 08:15:06 +0000

On March 19–20, 2026, the HClimRep group gathered in Hamburg for its third annual in-person meeting, hosted by the Hereon group at the International Maritime Museum. Partners from FZJ, AWI, KIT, and Hereon came together to review project progress, align future research directions, and strengthen collaboration across work packages. The meeting provided an important opportunity to assess current developments and define concrete decisions needed to guide the project through its next phase.

The first day focused on scientific and technical progress across core components of HClimRep. Updates covered advances in the stratospheric component led by KIT, progress in sub-seasonal and seasonal predictability by the FZJ group, the status of long-term rollout development presented by Hereon, and developments in the ocean component addressed by AWI members. Discussions also evaluated emerging competitors and clarified where the HClimRep AI–climate modelling approach provides distinctive advantages.

The second day shifted toward downstream applications, ongoing publications, and project infrastructure. These sessions aimed to ensure that technical developments translate efficiently into reproducible workflows and impactful scientific results. The meeting concluded with an open discussion on key challenges, future directions, and long-term ambitions.

The Hamburg meeting marked an important milestone in aligning scientific priorities, publications, and reinforcing collaboration as HClimRep moves toward its 2026-2027 goals.

Der Beitrag HClimRep annual in-person meeting 2026, Hamburg erschien zuerst auf HClimRep.

Internal Hackathon – AWI Bremerhaven

Agnes Lohs — Wed, 25 Mar 2026 08:03:20 +0000

On March 17–18, 2026, a small HClimRep Hackathon was held at the AWI partner’s offices in Bremerhaven. Five members of the HClimRep project – Ankit Patnala, Aleksei Koldunov, Jehangir Awan, Kacper Nowak, and Savvas Melidonis – gathered to work on climate‑forecasting and downscaling applications using tropospheric and oceanic data with the WeatherGenerator model.

The set‑up of the experiments progressed smoothly and was successful. In addition to the technical work, the team discussed the scientific agenda and publication strategy for the coming months, outlining the next steps in model development and planned outputs for the project.

Der Beitrag Internal Hackathon – AWI Bremerhaven erschien zuerst auf HClimRep.

HClimRep Presented at AlgoEarth & CESOC Workshop on Advancing AI Climate Models

Agnes Lohs — Mon, 16 Mar 2026 14:50:09 +0000

Cologne, 5th March, 2026

The HClimRep project was presented at the AlgoEarth & CESOC Workshop: Advancing AI Climate Models, held on 4–5 March 2026 at the University of Cologne. The workshop brought together international experts from research institutions, academia, and industry to discuss how artificial intelligence can move beyond short-term weather forecasting toward reliable climate modeling and long-term Earth system understanding. The event focused on key challenges in AI climate science, including limited observational data, complex feedback processes, and the need for new machine-learning architectures capable of representing coupled climate dynamics.

During the workshop, Dr. Savvas Melidonis (Jülich Supercomputing Centre, Forschungszentrum Jülich) presented “HClimRep: A Foundation Model for Capturing the Atmosphere, Ocean, and Sea Ice Interactions.” He explained how the project aims to develop an AI foundation model that learns the long-term interactions between atmosphere, ocean, stratosphere, and sea ice, addressing limitations of conventional physics-based climate models such as computational cost, structural biases, and coarse resolution. Furthermore, he stated that, by combining deep learning with high-performance computing (HPC), HClimRep seeks to bridge the gap between AI-based weather prediction and AI-driven climate projection.

The workshop, organized within the Algorithmic Earth System Science and CESOC frameworks, emphasized interdisciplinary collaboration across machine learning, algorithm design, and Earth system science. HClimRep’s contribution highlighted the growing role of AI foundation models in climate research and their potential to provide new insights into the Earth system while supporting evidence-based climate adaptation and policy decisions.

Der Beitrag HClimRep Presented at AlgoEarth & CESOC Workshop on Advancing AI Climate Models erschien zuerst auf HClimRep.

HClimRep annual in-person meeting 2025, Cologne, Germany

Florentine Weber — Thu, 11 Sep 2025 16:14:00 +0000

In early September, the HClimRep team gathered in Cologne for their annual project meeting. Nearly 20 researchers from different Helmholtz centers came together for two days of discussions on progress, challenges, and next steps.

A major focus of the meeting was the WeatherGenerator prototype, which integrates atmosphere, ocean, and stratosphere into a single framework. While this prototype forms the backbone of the project, it also highlights key challenges. Achieving stable forecasts over longer timescales remains a central bottleneck, reflecting inherent limits of current modelling approaches. Despite challenges around computing resources and data infrastructure, the meeting highlighted steady progress. The team has created new datasets that allow improved coupling of atmospheric and oceanic processes. Initial tests incorporating stratospheric chemistry tracers have been conducted, and alternative modeling strategies – such as diffusion-based approaches – are being explored in parallel. Strengthened evaluation methods and closer collaboration between work packages are further supporting advancement.

Looking ahead, the consortium agreed to focus efforts on refining the WeatherGenerator prototype, targeting a stable release by 2026. Upcoming priorities include achieving monthly-to-seasonal rollouts across all compartments, integrating climate forcing, and developing downscaling applications that link large-scale climate models with regional impacts. These efforts aim to provide both scientifically robust insights and practical tools for understanding and preparing for climate variability and change.

Der Beitrag HClimRep annual in-person meeting 2025, Cologne, Germany erschien zuerst auf HClimRep.

HClimRep annual in-person meeting 2024, Berlin, Germany

Eleonora Epp — Tue, 12 Nov 2024 09:44:36 +0000

The second HClimRep project meeting will take place at the premises of the Helmholtz Head Office in Berlin on Tuesday, November 26^th. We will review the current status of the ongoing HClimRep work packages and follow-up on our intense science discussions from the kickoff meeting in Cologne. On the day after the project meeting, a satellite meeting will be organized by the Synergy Unit of the Helmholtz Foundation Model Initiative, which will include final preparations for the Helmholtz Incubator workshop immediately afterwards. These will be excellent opportunities for networking and identifying synergies and potential collaborations among the HFMI pioneer projects.

Der Beitrag HClimRep annual in-person meeting 2024, Berlin, Germany erschien zuerst auf HClimRep.

HClimRep hackathon on 20. to 21. November 2024 at Cologne

Eleonora Epp — Thu, 24 Oct 2024 08:32:59 +0000

The HClimRep project members are organizing a 2-day hackathon event at the co-working space of “Die Wohngemeinschaft” in Cologne on November 20 to 21. The main objectives of the event are: 1.) Deepen knowledge about the Atmorep code and development for HClimRep project members who have already started to work with the code, 2.) Provide a general introduction to AtmoRep and train newcomers to obtain the code, data, and pretrained models and to run an AtmoRep application and training.
Twentyfive participants have registered and we are looking forward to two intense days of hacking and discussions.

Der Beitrag HClimRep hackathon on 20. to 21. November 2024 at Cologne erschien zuerst auf HClimRep.

Invited presentation at KAIST, Seoul, Korea

Eleonora Epp — Mon, 14 Oct 2024 08:31:00 +0000

Seoul, 25. September 2024

Martin Schultz gave an invited lecture at the KAIST Institute for AI in Seoul, Korea. The lecture summarized the rapid advancements of AI for weather and climate research and discussed the state-of-the-art as it was presented at the second workshop on “Large-scale deep learning for the Earth System” in Bonn, Germany, at the end of August. The talk inspired a two-hour lively and in-depth discussion with the very knowledgeable team members of Prof. Soo-Young, Lee.

Der Beitrag Invited presentation at KAIST, Seoul, Korea erschien zuerst auf HClimRep.

HClimeRep at the annual Helmholtz Assembly, Berlin

Eleonora Epp — Mon, 02 Sep 2024 11:27:21 +0000

Berlin, September 02, 2024

The HClimRep project was presented together with the other pilots from the Helmholtz Foundation Model Initiative at an interactive booth. HClimRep and AI development also appeared in the keynote lecture by the president Otmar Wiestler.

More information: https://www.helmholtz.de/ueber-uns/wer-wir-sind/presse-medien/veranstaltungen/jahrestagung-2024/

Der Beitrag HClimeRep at the annual Helmholtz Assembly, Berlin erschien zuerst auf HClimRep.