Preprints
https://doi.org/10.5194/hess-2021-511
https://doi.org/10.5194/hess-2021-511

  12 Oct 2021

12 Oct 2021

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

How can regime characteristics of catchments help in training of local and regional LSTM-based runoff models?

Reyhaneh Hashemi1, Pierre Brigode2,3, Pierre-André Garambois1, and Pierre Javelle1 Reyhaneh Hashemi et al.
  • 1Aix-Marseille Université, INRAE, UR RECOVER, Aix-en-Provence, France
  • 2Université Côte d’Azur, Observatoire de la Côte d’Azur, CNRS, IRD, Géoazur, Sophia-Antipolis, France
  • 3Université Paris-Saclay, INRAE, UR HYCAR, Antony, France

Abstract. In the field of Deep Learning, the long short-term memory (LSTM) networks lie in the category of recurrent neural network (RNN) architectures. The distinctive capability of the LSTM is learning non linear long term dependency structures. This makes the LSTM a good candidate for prediction tasks in non linear time dependent systems such as prediction of runoff in a catchment. In this study, we use a large sample of 740 gauged catchments with very diverse hydro-geo-climatic conditions across France. We present a regime classification based on three hydro-climatic indices to identify and classify catchments with similar hydrological behaviors. We do this because we aim to investigate how regime derived information can be used in training LSTM-based runoff models. The LSTM-based models that we investigate include local models trained on individual catchments as well as regional models trained on a group of catchments. In local training, for each regime, we identify the optimal lookback, i.e. the length of the sequence of past forcing data that the LSTM needs to work through. We then use this length in training regional models that differ in two aspects: 1) hydrological homogeneity of the catchments used in their training, 2) configuration of the static attributes used in their inputs. We examine how each of these aspects contributes to learning of the LSTM in regional training. At every step of this study, we benchmark performances of the LSTM against a conceptual model (GR4J) on both train and unseen data. We show that the optimal lookback is regime dependent and homogeneity of the train catchments in regional training has a more significant contribution to learning of the LSTM than the number of the train catchments.

Reyhaneh Hashemi et al.

Status: open (until 07 Dec 2021)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse

Reyhaneh Hashemi et al.

Reyhaneh Hashemi et al.

Viewed

Total article views: 315 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
252 59 4 315 4 0
  • HTML: 252
  • PDF: 59
  • XML: 4
  • Total: 315
  • BibTeX: 4
  • EndNote: 0
Views and downloads (calculated since 12 Oct 2021)
Cumulative views and downloads (calculated since 12 Oct 2021)

Viewed (geographical distribution)

Total article views: 244 (including HTML, PDF, and XML) Thereof 244 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 21 Oct 2021
Download
Short summary
Data driven LSTM models constitute a good candidate for prediction of runoff in a catchment. In this work, we used local LSTM models derived from individual catchments and regional models derived from a group of catchments. Our aim was to investigate how regime based information can be used for this class of models. We used a large sample of 740 catchments in the French climatic context. We identified three regime related factors that can make local and regional LSTM-based models more effective.