Preprints
https://doi.org/10.5194/hess-2022-56
https://doi.org/10.5194/hess-2022-56
 
10 Mar 2022
10 Mar 2022
Status: this preprint is currently under review for the journal HESS.

Improving hydrologic models for predictions and process understanding using Neural ODEs

Marvin Höge, Andreas Scheidegger, Marco Baity-Jesi, Carlo Albert, and Fabrizio Fenicia Marvin Höge et al.
  • Department of Systems Analysis, Integrated Assessment and Modelling, Eawag, Dübendorf, Switzerland

Abstract. Deep learning methods have frequently outperformed conceptual hydrologic models in rainfall-runoff modelling. Attempts of investigating the internals of such deep learning models are being made but traceability of model states and processes and their interrelations to model input and output is not yet fully given. Direct interpretability of mechanistic processes has always been considered as asset of conceptual models that helps to gain system understanding aside of predictability. We introduce hydrologic Neural Ordinary Differential Equation (ODE) models that perform as well as state-of-the-art deep learning methods in stream flow prediction while maintaining the ease of interpretability of conceptual hydrologic models. In Neural ODEs, internal processes that are represented in differential equations are substituted by neural networks. Therefore, Neural ODE models enable fusing deep learning with mechanistic modelling. We demonstrate the basin-specific predictive performance for several hundred catchments of the continental USA. For exemplary basins, we analyse the dynamics of states and processes learned by the model-internal neural networks. Finally, we discuss the potential of Neural ODE models in hydrology.

Marvin Höge et al.

Status: open (until 03 Jun 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'Comment on hess-2022-56', Juan Pablo Carbajal, 11 Apr 2022 reply
  • RC1: 'Comment on hess-2022-56', Andreas Wunsch, 07 May 2022 reply
  • RC2: 'Comment on hess-2022-56', Mr Miyuru Gunathilake, 20 May 2022 reply

Marvin Höge et al.

Marvin Höge et al.

Viewed

Total article views: 708 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
441 256 11 708 8 8
  • HTML: 441
  • PDF: 256
  • XML: 11
  • Total: 708
  • BibTeX: 8
  • EndNote: 8
Views and downloads (calculated since 10 Mar 2022)
Cumulative views and downloads (calculated since 10 Mar 2022)

Viewed (geographical distribution)

Total article views: 666 (including HTML, PDF, and XML) Thereof 666 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 26 May 2022
Download
Short summary
For stream flow predictions in hydrology, commonly two types of models are used: deep learning models (high predictive performance) and ODE-based conceptual hydrologic models (fully interpretable, encoding scientific assumptions). We introduce hydrologic Neural ODE models that fuse both approaches and have their benefits: We obtain state-of the-art predictive performance and gain insights into dynamics of model processes and states. We demonstrate the approach on a large real-world data set.