ECMWF, together with many partners, is extending AI-based weather forecasting beyond the atmosphere, towards representing the full Earth system, within the European Commission’s Destination Earth (DestinE) initiative. Our new page brings together the different machine learning (ML) Earth system components, including hydrology, waves, sea ice, ocean and land surface.  

By including insights from ECMWF scientists and combining blog articles and videos, this new page showcases how ECMWF together with Member States develop prototype ML models and the value they can bring for Earth system science.  

The different ML Earth components will be coupled with existing models, such as ECMWF’s Artificial Intelligence Forecasting System (AIFS) to form a flexible AI-driven Earth system model. These prototypes already deliver impressive results by leveraging state-of-the-art European datasets and EuroHPC supercomputers to move towards a fully ML-based Earth system model.  

Together, these ML components mark an important step towards a next-generation, fully AI-driven Earth system model within DestinE, enabling a more accurate and consistent representation of interactions across the Earth system. This is key to improving the accuracy, efficiency, and overall reliability of weather and climate predictions.