Articles | Volume 26, issue 12
Hydrol. Earth Syst. Sci., 26, 3079–3101, 2022
https://doi.org/10.5194/hess-26-3079-2022
Hydrol. Earth Syst. Sci., 26, 3079–3101, 2022
https://doi.org/10.5194/hess-26-3079-2022
Research article
20 Jun 2022
Research article | 20 Jun 2022

Hydrological concept formation inside long short-term memory (LSTM) networks

Thomas Lees et al.

Download

Interactive discussion

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'Comment on hess-2021-566', John Ding, 24 Nov 2021
    • CC2: 'Supplement on CC1', John Ding, 07 Dec 2021
    • AC1: 'Reply on CC1', Thomas Lees, 16 Feb 2022
  • RC1: 'Comment on hess-2021-566', Lukas Gudmundsson, 15 Dec 2021
    • AC2: 'Reply on RC1', Thomas Lees, 16 Feb 2022
  • RC2: 'Comment on hess-2021-566', Anonymous Referee #2, 17 Jan 2022
    • AC3: 'Reply on RC2', Thomas Lees, 16 Feb 2022

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision
ED: Publish subject to minor revisions (further review by editor) (07 Mar 2022) by Alexander Gruber
AR by Thomas Lees on behalf of the Authors (15 Mar 2022)  Author's response    Author's tracked changes    Manuscript
ED: Publish as is (04 Apr 2022) by Alexander Gruber
Download
Short summary
Despite the accuracy of deep learning rainfall-runoff models, we are currently uncertain of what these models have learned. In this study we explore the internals of one deep learning architecture and demonstrate that the model learns about intermediate hydrological stores of soil moisture and snow water, despite never having seen data about these processes during training. Therefore, we find evidence that the deep learning approach learns a physically realistic mapping from inputs to outputs.