Preprints
https://doi.org/10.5194/hess-2023-275
https://doi.org/10.5194/hess-2023-275
18 Jan 2024
 | 18 Jan 2024
Status: a revised version of this preprint is currently under review for the journal HESS.

HESS Opinions: Never train an LSTM on a single basin

Frederik Kratzert, Martin Gauch, Daniel Klotz, and Grey Nearing

Abstract. Machine learning (ML) has played an increasing role in the hydrological sciences. In particular, certain types of time series modeling strategies are popular for rainfall–runoff modeling. A large majority of studies that use this type of model do not follow best practices, and there is one mistake in particular that is common: training deep learning models on small, homogeneous data sets (i.e., data from one or a small number of watersheds). In this position paper, we show that Long Short Term Memory (LSTM) streamflow models are best when trained with a large amount of hydrologically diverse data.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Frederik Kratzert, Martin Gauch, Daniel Klotz, and Grey Nearing

Status: final response (author comments only)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on hess-2023-275', Marvin Höge, 05 Feb 2024
    • AC1: 'Reply on RC1', Frederik Kratzert, 28 Feb 2024
  • CC1: 'Comment on hess-2023-275', Sivarajah Mylevaganam, 05 Feb 2024
    • AC4: 'Reply on CC1', Frederik Kratzert, 28 Feb 2024
      • CC9: 'Reply on AC4', Sivarajah Mylevaganam, 29 Feb 2024
  • CC2: 'Comment on hess-2023-275', Sivarajah Mylevaganam, 06 Feb 2024
    • AC5: 'Reply on CC2', Frederik Kratzert, 28 Feb 2024
  • CC3: 'Comment on hess-2023-275', Sivarajah Mylevaganam, 06 Feb 2024
    • AC6: 'Reply on CC3', Frederik Kratzert, 28 Feb 2024
  • CC4: 'Comment on hess-2023-275', Sivarajah Mylevaganam, 07 Feb 2024
    • AC7: 'Reply on CC4', Frederik Kratzert, 28 Feb 2024
  • CC5: 'Comment on hess-2023-275', Sivarajah Mylevaganam, 09 Feb 2024
    • AC8: 'Reply on CC5', Frederik Kratzert, 28 Feb 2024
  • RC2: 'Comment on hess-2023-275', Markus Hrachowitz, 13 Feb 2024
    • AC11: 'Reply on RC2', Frederik Kratzert, 04 Mar 2024
  • RC3: 'Comment on hess-2023-275', Anonymous Referee #3, 15 Feb 2024
    • AC10: 'Reply on RC3', Frederik Kratzert, 28 Feb 2024
  • CC6: 'Comment on hess-2023-275', John Ding, 16 Feb 2024
    • AC2: 'Reply on CC6', Frederik Kratzert, 28 Feb 2024
  • RC4: 'Comment on hess-2023-275', Juliane Mai, 20 Feb 2024
    • AC3: 'Reply on RC4', Frederik Kratzert, 28 Feb 2024
  • CC7: 'Comment on hess-2023-275', Tam Nguyen, 27 Feb 2024
    • AC12: 'Reply on CC7', Frederik Kratzert, 04 Mar 2024
  • CC8: 'Comment on hess-2023-275', Sivarajah Mylevaganam, 28 Feb 2024
    • AC9: 'Reply on CC8', Frederik Kratzert, 28 Feb 2024
  • AC11: 'Reply on RC2', Frederik Kratzert, 04 Mar 2024
Frederik Kratzert, Martin Gauch, Daniel Klotz, and Grey Nearing

Data sets

Results and experimental data Frederik Kratzert, Martin Gauch, Daniel Klotz, and Grey Nearning https://doi.org/10.5281/zenodo.10139248

Model code and software

Code for analyzing model runs Frederik Kratzert, Martin Gauch, Daniel Klotz, and Grey Nearning https://github.com/kratzert/never-paper

Frederik Kratzert, Martin Gauch, Daniel Klotz, and Grey Nearing

Viewed

Total article views: 2,250 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
1,599 579 72 2,250 32 22
  • HTML: 1,599
  • PDF: 579
  • XML: 72
  • Total: 2,250
  • BibTeX: 32
  • EndNote: 22
Views and downloads (calculated since 18 Jan 2024)
Cumulative views and downloads (calculated since 18 Jan 2024)

Viewed (geographical distribution)

Total article views: 1,984 (including HTML, PDF, and XML) Thereof 1,984 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 

Cited

Latest update: 16 Jun 2024
Download
Short summary
Recently, a special type of neural network architecture became increasingly popular in hydrology literature. However, in most applications, this model was applied as a one-to-one replacement for hydrology models without adapting or rethinking the experimental setup. In this opinion paper, we show how this is almost always a bad decision and how using these kinds of models require the use of large-sample hydrology datasets.
Special issue