Regensburg 2025 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
DY: Fachverband Dynamik und Statistische Physik
DY 28: Poster: Machine Learning, Data Science
DY 28.5: Poster
Wednesday, March 19, 2025, 15:00–18:00, P4
Estimating parameters for a simple tipping model from complex Earth system model output — •Jonathan Krönke1,2, Jonathan F. Donges1, Johan Rockström1, Nils Bochow3, and Nico Wunderling1,2 — 1Earth Resilience Science Unit, Potsdam-Insitute for Climate Impact Research, Potsdam, Germany — 2Center for Critical Computational Studies, Goethe University Frankfurt, Frankfurt am Main, Germany — 3Department of Mathematics and Statistics, UiT - The Arctic University of Norway, Tromsø, Norway
The existence of large-scale tipping points - thresholds where small changes can trigger drastic, often irreversible shifts in the climate system - has been a major concern of climate science in the past two decades. The ability to evaluate tipping risks using computationally manageable models is crucial to assess the resilience of the climate system and also to identify safe global warming trajectories for tipping elements. Here, we present an approach to estimate parameters of a simple tipping model based on complex Earth system model output. We validate our results by reproducing simulations that have not been used in the training process and apply the model to major earth system tipping elements such as the Greenland Ice Sheet. A simple model that captures essential behaviour of complex earth system models provides an important step towards a tipping point emulator for extensive tipping risk analyses.
Keywords: Machine Learning; Climate Science; Tipping Points; Emulator; Dynamical Systems