Training the next generation of researchers is essential for accelerating scientific innovation and fostering interdisciplinary collaboration. With this in mind, CONCERTO contributed to the ELLIS Summer School 2025, held on 1 September 2025 in Jena, Germany, where early-career researchers and leading experts gathered to explore the latest developments in artificial intelligence and machine learning.

As part of the scientific programme, Amirpasha Mozaffari (Barcelona Supercomputing Center, BSC) presented the poster "Historical Reconstruction and Future Projection of Land Surface Boundary Conditions." The work demonstrates how machine learning can be applied to reconstruct historical land surface conditions and generate future projections, supporting improved representation of land processes within Earth system models. These advances contribute directly to CONCERTO's objective of enhancing terrestrial carbon cycle modelling through the integration of high-resolution Earth observation data and innovative AI methodologies.

The summer school provided an interdisciplinary environment for discussing emerging machine learning techniques and their application to real-world scientific challenges, while encouraging collaboration between researchers working across artificial intelligence, climate science and Earth observation. Through poster sessions and scientific exchanges, participants explored how data-driven approaches can help improve our understanding of complex environmental systems and support more reliable climate predictions.