Session GDB5

[Programme]

GDB5 | Forecasting the Unprecedented: Machine Learning vs. Physics in Climate Science and the EU’s Strategic Position

Forecasting the Unprecedented: Machine Learning vs. Physics in Climate Science and the EU’s Strategic Position

Convener: Athanasios Nenes | Co-convener: Ulas Im

Programme

| Fri, 08 May, 08:30–10:15 (CEST)

Room E1

Vimeo: GDB5 – Forecasting the Unprecedented: Machine Learning vs. Physics in Climate Science and the EU’s Strategic Position

As we approach the IPCC’s AR7 era, the landscape of climate prediction is more diverse—and contentious—than ever. Should we trust high-resolution, process-based models rooted in physical laws, or embrace the promise of machine learning, which some claim will soon surpass traditional approaches? With Earth observations and hybrid frameworks adding further complexity, a critical question emerges: can data-driven models anticipate a future climate that will be fundamentally different from anything in the observational record? This debate is urgent, as society demands actionable guidance on climate risks and tipping points.
At the same time, global leadership in climate science is shifting. With recent political developments in the US, the European Union faces both a challenge and an opportunity to shape the scientific and policy agenda. What should Europe’s role be in steering the next generation of climate modeling and ensuring robust, transparent advice for decision-makers? This session brings together leading voices from science, policy, and technology to debate the future of climate prediction, the limits of machine learning, and the responsibilities of the EU in a rapidly changing world.