Dr. Ralf Loritz

KIT-Junior Research Leader & KIT Associate Fellow, Institute for Water and Environment – Hydrology (IWU/HYD), KIT

The lecture will demonstrate how machine learning (ML) is transforming hydrology – and why these technologies have become a key tool for predicting floods, droughts, and other hydrological extremes. 

Hydrology deals with the water cycle in all its facets – from precipitation to storage in the soil to runoff into rivers and lakes. Mathematical and physical models have been used for decades to understand these processes and predict future events. They represent the well-known laws of hydrology – such as infiltration, evaporation, and runoff formation – through equations, thereby enabling a quantitative description of water movement across the landscape.
But are new, data-driven approaches superior to these models? What are the respective advantages and disadvantages of each method? These questions are at the heart of the lecture.

Using examples from current research projects, such as the BMFTR-funded collaborative project KI-HopE, the talk will also demonstrate how current concepts are being implemented in practice – for instance, in the development of a Germany-wide, machine learning-based flood forecasting system in cooperation with the Deutscher Wetterdienst and the flood forecasting centers. 

This presentation invites us to rethink hydrology as a data science discipline – how can an understanding of physics and artificial intelligence help us better identify the risks of flooding and drought and address them earlier?

The presentation will also introduce the KIT Graduate School of Computational and Data Science.

Angela Hühnerfuß, M.A.

Managing Director, KIT Graduate School Computational and Data Science, KIT

The KIT Graduate School Computational and Data Science | KCDS at KIT Center MathSEE is an English-language graduate school for all doctoral researchers at KIT who are working on interdisciplinary, data- and model-driven research projects.

What sets KCDS apart is the tandem principle: Each doctoral candidate has one advisor from the mathematical sciences with a focus on methodology and one from the SEE discipline (natural sciences, economics, and engineering) with a focus on application. The interdisciplinary training program at KCDS offers doctoral candidates the opportunity to pursue targeted professional and interdisciplinary development, expand their network, and organize their own projects and events.

External guests may attend individual sessions provided there is sufficient capacity. In this case, please send an email request to dataliteracy∂forum.kit.edu. You may attend up to two sessions free of charge.

If there is significant interest in participating in the lecture series beyond that, you may attend as a registered guest student starting with the third session.