The IMITATE project, led by the University of Leicester under ESA’s EO4Society programme, explored how machine learning can be used to emulate and enhance physical land surface models of terrestrial ecosystems. Focusing on Europe and the carbon cycle, the project developed AI-based “emulators” of the JULES Earth System Model and compared their outputs with both model simulations and Earth Observation data. IMITATE aimed to improve understanding of key land-atmosphere processes and assess how well data-driven relationships align with established Earth system models. The project also sought to generate new EO-based products, including estimates of ecosystem productivity, to support climate and environmental research.