Impact of training dataset size and its hydrometeorological typology on LSTM performance for rainfall-runoff modeling: a case study of the Severn river - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Impact of training dataset size and its hydrometeorological typology on LSTM performance for rainfall-runoff modeling: a case study of the Severn river

IMPACT DE LA TAILLE DU JEU DE DONNÉES D'ENTRAÎNEMENT ET DE SA TYPOLOGIE HYDROMÉTÉOROLOGIQUE SUR LES PERFORMANCES DES LSTM POUR LA MODÉLISATION DES ÉCOULEMENTS PLUVIAUX : UNE ÉTUDE DE CAS DE LA RIVIÈRE SEVERN

Résumé

Accurate discharge prediction in hydrological forecasting relies on robust modeling. This study investigates the Long Short-Term Memory (LSTM) model's performance, focusing on training dataset size and hydrometeorological patterns. Convolutional Neural Networks (CNNs) and Artificial Neural Networks (ANNs) are also considered for spatial and temporal dependencies. Data is drawn from the CAMELS-GB dataset (1975-2015, Saxons Lode, UK). Results show that LSTM performance varies, with years surrounding high water events (like 2004) performing poorly in training, and struggles in validation. Training with one year yields 23.03% NSE values above 0.7, but using three consecutive years improves this to 84.42%. Typological differences also affect model performance. This study reveals LSTM sensitivity to training periods, aiding the optimization of training duration for better discharge prediction accuracy. Future research will delve into year selection within typological clusters.
Fichier principal
Vignette du fichier
simhydro2023_shorten.pdf (875.03 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04375806 , version 1 (05-01-2024)

Identifiants

  • HAL Id : hal-04375806 , version 1

Citer

Nadia Skifa, Fadil Boodoo, Carole Delenne, Renaud Hostache, Morgan Abily. Impact of training dataset size and its hydrometeorological typology on LSTM performance for rainfall-runoff modeling: a case study of the Severn river. SimHydro conferences, Société Hydrotechnique de France (SHF); the Association Française de Mécanique (AFM); the Environmental & Water Resources Institute (EWRI); the International Association for Hydro-Environment Engineering and Research (IAHR), Nov 2023, Chatou, France. ⟨hal-04375806⟩
47 Consultations
33 Téléchargements

Partager

Gmail Facebook X LinkedIn More