Climate models can predict where the summer will be particularly hot in Europe or whether an unusually strong El Niño will develop over the Pacific. At a global level, they already successfully depict the consequences of various future scenarios. However, predictions about climate change are only useful if the earth system models are as precise as possible.
At climate conferences in Japan and Austria, two researchers from Climate QML will therefore demonstrate how the DLR team uses machine learning and quantum computers to reduce inaccuracies in data analyses precisely where turbulent chaos prevails. At the DLR Institute of Atmospheric Physics , they are researching how quantum machine learning (QML) can be used to model climate more accurately and quickly, also in order to improve technology assessments and mitigation recommendations.
Machine learning improves climate models
Modelling and predicting atmospheric turbulence is particularly challenging: turbulence is inherently chaotic and behaves very differently, especially in the atmospheric boundary layer. Climate QML is investigating how climate models can be improved with the help of machine learning and quantum computers. Mierk Schwabe recently presented the latest results at the CMIP Community Workshop 2026 in Kyoto, Japan.
More precise analysis with quantum computers
Klim-QML will also be represented at the General Assembly of the European Geosciences Union (EGU26). At this international conference in Vienna from 3 to 8 May, Lena Dogra will speak about data analysis of turbulence with quantum computers. Recently, hybrid models based on machine learning have successfully improved the parameterisations in the studies. Because they are data-driven, they can analyse the empirical data behind them more precisely.
However, improved climate models, which are intended to make more accurate predictions at a regional and local level, require high computing power. Mierk Schwabe, Lena Dogra and the team at the DLR Institute of Atmospheric Physics are investigating how quantum computers can help here. As a reference, they are comparing and evaluating the optimised models with real weather phenomena in order to enable more accurate predictions of regional and local changes. However, such investigations not only lead to more precise forecasts: optimised climate models also help to understand the effects of various technologies from the aerospace, energy and transport sectors.
