Meteorologist, Physicist or similar (f/m/x)

Deep moist convection constitutes one of the most important atmospheric transport mechanisms and is associated with several extreme-weather events (hail, heavy precipitation, wind gusts, …). As a consequence of climate change, the severity of these events is expected to increase in the future. Spatiotemporally highly resolved short-term predictions of deep moist convection are an essential tool to limit significant social and economic damage. Applications range from automated transportation to optimal routing of air traffic.

The objective of this PhD thesis is the development of a more accurate, robust and physically consistent short-term forecasting method for deep moist convection. The starting point will be a prototypical prediction method currently under development, in which we combine physical model building with deep learning. In a first step, the existing model shall be extended to the use of data from the new Meteosat Third Generation (MTG) satellite mission, of which we are among the first users. In a second step and in close collaboration with the DLR Institute of Data Science in Jena, uncertainties and explainability of the model will be established. On the basis of these analyses, potentials for improvement shall be identified and implemented. This work may benefit from concepts investigated in an ongoing PhD thesis and shall explicitly exploit synergies. In addition to thorough validation of the model, case studies in aviation demonstrators are integral part of the project.