From Topology to Dynamics: Advancing River Network Representation in Physical and Machine Learning Models
From Topology to Dynamics: Advancing River Network Representation in Physical and Machine Learning Models
Session ID#: 281933
Session Description:
River networks are fundamental to the hydrological cycle, yet their representation in both numerical and data-driven models remains a key source of uncertainty across scales. This session invites contributions that advance how river networks are conceptualized, structured, parameterized, and monitored — spanning innovations in network topology, connectivity, and routing that are applicable to physics-based, machine learning, and/or hybrid modeling approaches. We welcome submissions from catchment to global scales and across disciplinary boundaries, including hydrology, Earth system modeling, and geospatial data science. Topics may include novel network representations and parameterizations, graph-based and deep learning approaches, multi-scale routing schemes, solutions to the network-to-network and network-to-gauge site mapping problems, and integration of river networks within land surface and Earth system models.
Co-Sponsor(s):
- EP - Earth and Planetary Surface Processes
Index Terms:
1847 Modeling [HYDROLOGY]
1856 River channels [HYDROLOGY]
1860 Streamflow [HYDROLOGY]
1942 Machine learning [INFORMATICS]
Primary Convener: Louise J. Slater, University of Oxford, School of Geography and the Environment, Oxford, United Kingdom
Conveners: Michel Wortmann, European Center for Medium-Range Weather Forecasts, Reading, United Kingdom, Simon Moulds, University of Edinburgh, School of Geosciences, Edinburgh, United Kingdom, Peirong Lin, Peking University, Institute of Remote Sensing and GIS, School of Earth and Space Sciences, Beijing, China and Frederik Kratzert, Google Research, Vienna, Austria
See more of: Hydrology