High-performance computing in the hydrosciences

This paper provides an overview of High Performance Computing (HPC) in the hydrosciences. It begins with the development of HPC architectures and performance as well as with the expansion of application fields over the last decades. Subsequently, four HPC application cases are presented. The first case considers high-resolution simulation of 3D two-phase flow, transport and mass transfer processes in sewer systems, addressing problems of odour and corrosion. In the second case, high-resolution rainfall-runoff modelling with a 2D shallow water model is investigated and applied to the entire city of Berlin, Germany, to detect critical areas. The third case deals with simulation-optimisation of 2D river flow and water quality in a section of the Nile River and highlights the urgent necessity of automatic calibration. In the fourth case, a fast prediction method for heavy precipitation-induced runoff in a district of Berlin using a deep learning approach is developed, enabling early warning. Overall, these application cases demonstrate the urgent necessity and the wide range of HPC in the hydrosciences and its ability to solve problems which could not have been addressed without it.