Exploring the Potential of Machine Learning for Improving Operational Hydrological Forecasting and Prediction

Doctoral Researcher
Name Role at KCDS
KCDS Supervisors
Name Role at KCDS
Deputy Scientific Speaker, SEE Supervisor
MATH Supervisor, member of the Steering Committee

Abstract

Machine Learning (ML) methods has been revolutionizing hydrological modeling and ML-based streamflow prediction is one of the main topics in the field. In close cooperation with the Landesanstalt für Umwelt Baden-Württemberg (LUBW), this project aims to systematically explore how ML methods can improve operational hydrological forecasting, especially for floods, along the entire forecasting chain, to further develop existing ML methods, and to improve ML-based uncertainty quantification. The project’s outcomes will provide theoretical advancements of ML and uncertainty quantification, and also help to substantially improve and accelerate the operational hydrological forecasting practice in Baden-Württemberg.