Our planet is warming faster than at any point in recorded history. As a result of this warming, people around the world are experiencing heatwaves, droughts, floods and wildfires affecting food, water, health and economies worldwide.Identifying the impacts of global warming as well as extreme event forecasting thus are crucial research directions for climate scientists. Two major challenges however remain:
1. Climate downscaling: Many applications in e.g., agriculture, environmental conservation and hydrology management require fine-grained data on a small area. A major challenge thus is climate downscaling to obtain fine-grained climate details from coarse-grained inputs.
2. Extreme-event forecasting: Due to extreme events being rare, data to learn from remains scarce. Additionally, they often arise as a result of complex, chaotic interactions which are hard to study. Understanding extreme precipitation and temperature events thus remains a major challenge in climate research. While traditional numerical methods provide a solid starting point, machine learning methods have shown to accelerate discovery and solution finding in climate research. We are therefore developing new machine learning methods for climate research with a particular focus on climate downscaling and extreme-event forecasting.
In particular, we will focus on Fourier Neural Operators that are mesh-free and enable zero-shot super-resolution predictions building on our recent work [1].
Sample of our MS students/interns
Next one maybe you?!
Josep Peiro Ramos (First author of ICLR submission)
Ola Karlstad (Co-author of ICLR submission)
Frida Marie Engøy Westby (First author of ICLRW 2026 paper)
Preben Nicholai Castberg (Co-author of ICLR submission)
Brage Eilertsen (First author of AAAI 26 paper)
Røskva Bjørgfinsdóttir (Co-author of AAAI 26 paper)
Iman Tabrizian (Co-first author of NeurIPS paper)
[1] Vahidreza Jahanmard, Robinson Hordoir, Ali Ramezani‐Kebrya, Artu Ellmann, Nicole Delpeche‐Ellmann, "Principled Fourier neural operators for high‐resolution regional ocean modeling", Journal of Geophysical Research: Machine Learning and Computation 2026.