Start of funding 01.01.2021

AI-augmented Cloud Tomography for next-generation Remote Sensing of Cloud Properties to Improve Short-Term Solar Energy Forecasting

Prof. Dr. Bernhard Mayer
Ludwig-Maximilians-University of Munich
Lehrstuhl für Experimentelle Meteorologie

Prof. Dr. Katherine L. Bouman
California Institute of Technology, Pasadena (CALTECH)
Department of Computing and Mathematical Sciences



Clouds are a key component of the Earth’s climate system, and artificial intelligence (AI) has the potential of significantly advancing observation techniques relating to cloud properties. This project aims at developing a novel tomographic remote sensing method similar to Computed Tomography in medical imaging. We will develop an AI-augmented retrieval method for reconstruction of 3D cloud properties from multi-angle observations. In traditional retrievals, clouds are simplified as one-dimensional, horizontally homogeneous cloud layers. This assumption causes a significant bias for convective clouds, which make up a large fraction of global cloud coverage. Thus, advanced methods for investigation of cloud properties are needed to improve short-term forecasting for solar energy harvesting. Furthermore, long-term monitoring of cloud properties via satellite will improve our understanding of the clouds’ role in the Earth’s energy budget.