Research highlights

We combine machine learning/AI, numerical models and Earth observations to address key challenges in the climate and environmental sciences. Below we link to recent research highlights that have been published in leading field-specific and interdisciplinary journals.

For example,

  • using satellite data, we develop novel machine learning approaches to reduce uncertainties in global climate change projections.
  • we are working on new types of hybrid climate models, which include both classical numerical and novel machine learning components. Hybrid models could not only improve individual process representations but also lower the computational costs of climate models run on supercomputers.
  • we use machine learning to improve our understanding of the complex, highly-coupled climate system. In particular, we are interested in data-driven causal discovery and explainable AI.
  • we are working on machine learning techniques for better and more affordable measurements of the Earth system.
Model uncertainty
AI and satellite data
A supercomputer KIT
AI for climate modelling
Causal discovery PJN
AI for inference
Sensors AirPublic Ltd
AI for sensors