When I started as Professor for Computational Analytics at the University of Bonn in October 2019, the field of Geodesy was not yet one of my research interests, although my research has always been application-oriented. Soon after, I was involved in the formation of a group of scientists, that felt addressed by a DFG call with the objective of (1) closely link fundamental research in AI with a new partner field and (2) integrate the topic onto a local long-term research profile. Building on the long-standing close relationship between the Universities of Bonn and Cologne within the field of Geoscience, Geodesy was identified as a promising field of application for AI, so that both objectives of the call were met. In February 2020, Anne Driemel, Heiko Röglin, Jan-Henrik Haunert, Jürgen Kusche and I jointly with colleagues from the University of Cologne, Melanie Schmidt and Christian Sohler, started planning the proposal to the DFG. In order to strengthen our joint preliminary work in the field of Geodesy, it was very helpful to receive financial support from TRA Modelling by funding a Ph.D. position for 15 months. Besides financial support, also the positive feedback on our TRA application really helped us to move the project forward.
To our delight, the DFG decided in June 2022 to fund our AI research unit FOR 5361 “AlgoForGe: Algorithmic Data Analytics for Geodesy”, in which we study algorithmic challenges of fundamental AI problems in Geodesy, and develop new algorithmic data analysis methods and tools. Since our common preliminary work for the DFG proposal has been extra-ordinarily successful, the three founders of CESOC (Susanne Crewell, Astrid Kiendler-Scharr, and Jürgen Kusche) asked us in November 2020 to join. CESOC is a joint virtual institute of the Universities of Bonn, Cologne, and FZ Jülich, which has been founded in October 2020 to support the successful application of Bonn to host the new location for the European Centre for Medium-Range- Weather Forecasts (ECMWF). CESOC focuses on physical, mathematical and biogeochemical modeling of the coupled Earth system in view of the development of a digital Earth model, and aims at close ties between geosciences and methods from Computer Science and Mathematics.
Overall, these developments were helpful for a planned joint Cluster initiative, which unfortunately was not invited to submit a full proposal in 2024. We are nevertheless confident that the new field of Algorithmic Earth System Science planned there will be established on the long run and the cooperation of the Universities Cologne, Bonn, FZ Jülich, and the ECMWF will be continued successfully. For example, Bonn and Cologne are jointly planning for a semester in Algorithmic Earth Science in winter term of 2025/26. Our overall aim is to better integrate concepts from algorithmics in Theoretical Computer Science and Earth Observation in order to significantly improve our understanding and predictive capabilities of the rapidly changing Earth system at regional-to-local scales.