Introducing HydroPy, an updated global hydrology model rewritten in Python

Global hydrological models (GHM) are a useful tool to investigate the water cycle, to evaluate its sensitivity towards systematic changes, e.g. human impacts, and to project future conditions in river catchments for varying scenarios. They have been successfully applied for decades and there is still room for improvement.

Recently, we revised the Max Planck Institute for Meteorology’s Hydrology model (MPI-HM), which is an established GHM that was used in multiple case studies and inter-comparison projects. While still performing well, its source code (mainly Fortran77) has become increasingly difficult to maintain, thus hampering the implementation of new processes. For this reason, the model was rewritten from scratch based on the MPI-HM process formulations. The new model is mainly written in Python, thereby taking advantage of the highly optimized numpy and xarray libraries, and, hence, is aptly renamed to HydroPy. Using the original formulations, we make sure to preserve or even improve the old model’s skill while the switch to Python allows for much easier debugging and interactive model development.

In our presentation, we will evaluate the performance of the new HydroPy model and demonstrate its skill to simulate river discharge. Furthermore, we compare HydroPy to its predecessor MPI-HM and discuss the reasons of differences between their results.