Using artificial intelligence for better understanding the Earth system
European Union funds the international research project AI4PEX to further improve Earth system models and thus scientific predictions of climate change. Participating scientists from 9 countries met at the end of May 2024 to launch the project at the Max Planck Institute for Biogeochemistry in Jena, which is leading the project.
Earth system models simulate the major physical and biogeochemical cycles and serve as a basis for projections of Earth’s future climate. These models simulate interactions between important geophysical and geochemical processes in the climate system, including the atmosphere, the land surface, the oceans and all bodies of water, as well as human-induced activities and changes such as greenhouse gas emissions. As such, the appropriate representation of such interactions and feedbacks is key in their development.
The overarching goal of AI4PEX is to further improve the processes underpinning Earth system feedbacks. To this end, AI4PEX researchers are bringing Earth observations and models together using artificial intelligence and machine learning for modeling and analyzing present and future projections of the Earth system. With this approach AI4PEX aims for an improved representation of key processes in the Earth system, as well as the representation of extreme events, in complex and physically-based models. The project uses observation data and models of the land surface, the atmosphere and the oceans. Ultimately, AI4PEX aims at a better understanding of the Earth system while paving the way for more reliable climate projections on global and regional scales.
The project team got together to officially launch AI4PEX in the Max Planck Institute for Biogeochemistry (MPI-BGC), in Jena. During the three-day meeting, the team embarked on discussing complex issues such as cloud feedbacks in the atmosphere, heat and carbon exchange in the ocean, the terrestrial carbon cycle, Earth system models, the representation of extreme weather events in climate models, machine learning and artificial intelligence, as well as hybrid models. In addition to these technical topics, one day was reserved to train the new staff in the use of the EuroHPC supercomputer LUMI, which is used for data processing in the project, and to introduce them to the ESMVal, a diagnostic and performance metrics tool for the routine evaluation of Earth system models.
The project is funded by the European Union's Horizon Europe program, UK Research and Innovation, and the Swiss State Secretariat for Education, Research and Innovation with a project duration of 4 years. In addition to the Max Planck Institute for Biogeochemistry in Jena, which is coordinating the project, 18 other European institutions from a total of nine countries are involved in AI4PEX, namely: National Center for Scientific Research (Gif sur Yvette, France), CSC - IT Center for Science LTD. (Espoo, Finland), German Aerospace Center (Oberpfaffenhofen, Germany), Technical University of Denmark (Copenhagen, Denmark), Eberhard Karls University Tübingen, Institut Mines-Télécom (Palaiseau, France), National Center for Meteorological Research of Meteo France (Saint Mande Cedex, France), Swedish Meteorological and Hydrological Institute (Norrköpping, Sweden), University of Leipzig, University of Lund (Sweden), University of Valencia (Italy), Flemish Institute for the Sea (Ostend, Belgium), University of Hamburg, Swiss Federal Institute of Technology Zurich (Switzerland), University of Lausanne (Switzerland), Met Office (Exeter, UK), University of Leeds (UK), and University of Reading (UK).
AI4PEX is led by Dr. Nuno Carvalhais (scientific project coordinator) and Dr. Nicole Börner (project manager), both at the MPI-BGC on Beutenberg Campus in Jena.
AI4PEX is led by Dr. Nuno Carvalhais (scientific project coordinator) and Dr. Nicole Börner (project manager), both at the MPI-BGC on Beutenberg Campus in Jena.