Preprints
https://doi.org/10.5194/hess-2021-371
https://doi.org/10.5194/hess-2021-371

  24 Aug 2021

24 Aug 2021

Review status: this preprint is currently under review for the journal HESS.

Impact of Spatial Distribution Information of Rainfall in Runoff Simulation Using Deep-Learning Methods

Yang Wang and Hassan A. Karimi Yang Wang and Hassan A. Karimi
  • Geoinformatics Laboratory, School of Computing and Information, University of Pittsburgh, 135 N Bellefield Ave, Pittsburgh, PA 15213, USA

Abstract. Rainfall-runoff modelling is of great importance for flood forecast and water management. Hydrological modelling is the traditional and commonly used approach for rainfall-runoff modelling. In recent years, with the development of artificial intelligence technology, deep learning models, such as the long short-term memory (LSTM) model, are increasingly applied to rainfall-runoff modelling. However, current works do not consider the effect of rainfall spatial distribution information on the results, and the same look-back window is applied to all the inputs. Focusing on two catchments from the CAMELS dataset, this study first analyzed and compared the effects of basin mean rainfall and spatially distributed rainfall data on the LSTM models under different look-back windows (7, 15, 30, 180, 365 days). Then the LSTM+1D CNN model was proposed to simulate the situation of short-term look-back windows (3, 10 days) for rainfall combined with the long-term look-back windows (30, 180, 365 days) for other input features. The models were evaluated using the Nash Sutcliffe efficiency coefficient, root mean square error, and error of peak discharge. The results demonstrate the great potential of deep learning models for rainfall runoff simulation. Adding the spatial distribution information of rainfall can improve the simulation results of the LSTM models, and this improvement is more evident under the condition of short look-back windows. The results of the proposed LSTM+1D CNN are comparable to those of the LSTM model driven by basin mean rainfall data and slightly worse than those of spatially distributed rainfall data for corresponding look-back windows. The proposed LSTM+1D CNN provides new insights for runoff simulation by combining short-term spatial distributed rainfall data with long-term runoff data, especially for catchments where long-term rainfall records are absent.

Yang Wang and Hassan A. Karimi

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-2021-371', Anonymous Referee #1, 05 Oct 2021
    • AC1: 'Reply on RC1', Yang Wang, 09 Oct 2021
    • AC2: 'Reply on RC1', Yang Wang, 11 Oct 2021
  • CC1: 'Comment on hess-2021-371', Qingtai Qiu, 24 Oct 2021
    • AC3: 'Reply on CC1', Yang Wang, 27 Oct 2021
  • RC2: 'Comment on hess-2021-371', Anonymous Referee #2, 12 Nov 2021
    • AC4: 'Reply on RC2', Yang Wang, 19 Nov 2021
  • EC1: 'Comment on hess-2021-371', Dimitri Solomatine, 30 Nov 2021

Yang Wang and Hassan A. Karimi

Yang Wang and Hassan A. Karimi

Viewed

Total article views: 602 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
441 143 18 602 1 0
  • HTML: 441
  • PDF: 143
  • XML: 18
  • Total: 602
  • BibTeX: 1
  • EndNote: 0
Views and downloads (calculated since 24 Aug 2021)
Cumulative views and downloads (calculated since 24 Aug 2021)

Viewed (geographical distribution)

Total article views: 505 (including HTML, PDF, and XML) Thereof 505 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 02 Dec 2021
Download
Short summary
Different look-back windows should be explored to obtain the optimal results when using LSTM for rainfall-runoff simulations. Adding the spatial distribution information of rainfall can improve the results of the LSTM model, especially peak discharge. The results of our proposed LSTM+1D CNN on 'n time step output' and 'one time step output' are comparable to those of the LSTM model driven by basin mean rainfall data, and slightly worse than those of spatially distributed rainfall data.