Preprints
https://doi.org/10.5194/hess-2021-127
https://doi.org/10.5194/hess-2021-127

  12 Mar 2021

12 Mar 2021

Review status: a revised version of this preprint was accepted for the journal HESS and is expected to appear here in due course.

Benchmarking Data-Driven Rainfall-Runoff Models in Great Britain: A comparison of LSTM-based models with four lumped conceptual models

Thomas Lees1, Marcus Buechel1, Bailey Anderson1, Louise Slater1, Steven Reece2, Gemma Coxon3, and Simon J. Dadson1,4 Thomas Lees et al.
  • 1School of Geography and the Environment, University of Oxford, South Parks Road, Oxford, United Kingdom, OX1 3QY
  • 2Department of Engineering, University of Oxford, Oxford, United Kingdom
  • 3Geographical Sciences, University of Bristol, Bristol, United Kingdom
  • 4UK Centre for Ecology and Hydrology, Maclean Building, Crowmarsh Gifford, Wallingford, United Kingdom, OX10 8BB

Abstract. Long short-term memory models (LSTMs) are recurrent neural networks from the emerging field of Deep Learning (DL), which have shown recent promise when predicting time-series especially when data are abundant. Rainfall-runoff modelling presents a challenge, yet accurate hydrological models are vital for flood forecasting, hazard impact assessment, and to assess the potential effects of climate change on floods and water resources. In this study, we compare the performance of two DL-based models, a LSTM and an Entity Aware LSTM (EA LSTM). The DL models were trained using a newly published data set, CAMELS-GB, for a sample of 518 catchments across Great Britain. To identify spatial and seasonal patterns in model performance, we compare the DL models against benchmark outputs from four lumped conceptual models recently configured for rainfall-runoff modelling in Great Britain. Our findings show that the LSTM models simulate discharge with consistently high model performance scores, including in catchments typically considered difficult to model. The LSTM achieves a mean catchment NSE of 0.88 (0.86 for the EALSTM), which represents a performance improvement of 10 %–16 % compared with the benchmark conceptual models. Seasonal and spatial patterns indicate that the largest performance improvement relative to the benchmark is in the drier summer months and in drier catchments in the South East of England. By comparing LSTMs with conceptual models, we diagnose possible reasons for their different performance. We suggest that LSTMs offer useful predictive capability for rainfall-runoff modelling in Great Britain and elsewhere and note their value to support process understanding in locations where processes are less well understood.

Thomas Lees et al.

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'Comment on hess-2021-127', Alison Kay, 16 Mar 2021
    • CC2: 'Reply on CC1', Thomas Lees, 16 Mar 2021
      • CC3: 'Reply on CC2', Alison Kay, 16 Mar 2021
      • AC4: 'Reply on CC2', Thomas Lees, 14 Jun 2021
  • RC1: 'Comment on hess-2021-127', Anonymous Referee #1, 29 Mar 2021
    • AC1: 'Reply on RC1', Thomas Lees, 14 Jun 2021
  • RC2: 'Comment on hess-2021-127', Anonymous Referee #2, 06 Apr 2021
    • AC2: 'Reply on RC2', Thomas Lees, 14 Jun 2021
  • RC3: 'Comment on hess-2021-127', Anonymous Referee #3, 18 May 2021
    • AC3: 'Reply on RC3', Thomas Lees, 14 Jun 2021
  • AC5: 'Overview of Proposed Manuscript Changes based on Reviewer Comments', Thomas Lees, 14 Jun 2021
  • AC6: 'Overview of Proposed Manuscript Changes based on Reviewer Comments', Thomas Lees, 14 Jun 2021

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'Comment on hess-2021-127', Alison Kay, 16 Mar 2021
    • CC2: 'Reply on CC1', Thomas Lees, 16 Mar 2021
      • CC3: 'Reply on CC2', Alison Kay, 16 Mar 2021
      • AC4: 'Reply on CC2', Thomas Lees, 14 Jun 2021
  • RC1: 'Comment on hess-2021-127', Anonymous Referee #1, 29 Mar 2021
    • AC1: 'Reply on RC1', Thomas Lees, 14 Jun 2021
  • RC2: 'Comment on hess-2021-127', Anonymous Referee #2, 06 Apr 2021
    • AC2: 'Reply on RC2', Thomas Lees, 14 Jun 2021
  • RC3: 'Comment on hess-2021-127', Anonymous Referee #3, 18 May 2021
    • AC3: 'Reply on RC3', Thomas Lees, 14 Jun 2021
  • AC5: 'Overview of Proposed Manuscript Changes based on Reviewer Comments', Thomas Lees, 14 Jun 2021
  • AC6: 'Overview of Proposed Manuscript Changes based on Reviewer Comments', Thomas Lees, 14 Jun 2021

Thomas Lees et al.

Thomas Lees et al.

Viewed

Total article views: 11,281 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
5,869 5,384 28 11,281 10 3
  • HTML: 5,869
  • PDF: 5,384
  • XML: 28
  • Total: 11,281
  • BibTeX: 10
  • EndNote: 3
Views and downloads (calculated since 12 Mar 2021)
Cumulative views and downloads (calculated since 12 Mar 2021)

Viewed (geographical distribution)

Total article views: 10,142 (including HTML, PDF, and XML) Thereof 10,142 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 16 Oct 2021
Download
Short summary
We used deep learning models to simulate the amount of water travelling through a river channel (discharge), based on the rainfall, temperature and potential evaporation in the previous days. We compared the performance of these deep learning models to four models that are currently used by hydrologists and engineers. We tested how these different methods simulate discharge at 518 points around Great Britain. Ultimately, we found that the deep learning models produced more accurate simulations.