Technical note: An approach for handling multiple temporal frequencies with different input dimensions using a single LSTM cell

Acuña Espinoza, Eduardo; Kratzert, Frederik; Klotz, Daniel; Gauch, Martin; Álvarez Chaves, Manuel; Loritz, Ralf; Ehret, Uwe

doi:https://doi.org/10.5194/hess-29-1749-2025

Articles | Volume 29, issue 6

https://doi.org/10.5194/hess-29-1749-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/hess-29-1749-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 29, issue 6

Technical note

|

26 Mar 2025

Technical note |

| 26 Mar 2025

Technical note: An approach for handling multiple temporal frequencies with different input dimensions using a single LSTM cell

Eduardo Acuña Espinoza, Frederik Kratzert, Daniel Klotz, Martin Gauch, Manuel Álvarez Chaves, Ralf Loritz, and Uwe Ehret

Data sets

Models and Predictions for "Rainfall-Runoff Prediction at Multiple Timescales with a Single Long Short-Term Memory Network" M. Gauch et al. https://doi.org/10.5281/zenodo.4095485

Data for "Rainfall-Runoff Prediction at Multiple Timescales with a Single Long Short-Term Memory Network" M. Gauch et al. https://doi.org/10.5281/zenodo.4072701

CAMELS: Catchment Attributes 290 and MEteorology for Large-sample Studies. Version 1.2 A. Newman et al. https://doi.org/10.5065/D6G73C3Q

CAMELS Extended Maurer Forcing Data F. Kratzert https://doi.org/10.4211/hs.17c896843cf940339c3c3496d0c1c077

Model code and software

An approach for handling multiple temporal frequencies with different input dimensions using a single LSTM cell Eduardo Acuna Espinoza https://doi.org/10.5281/zenodo.14780059

Short summary

Long short-term memory (LSTM) networks have demonstrated state-of-the-art performance for rainfall-runoff hydrological modelling. However, most studies focus on predictions at a daily scale, limiting the benefits of sub-daily (e.g. hourly) predictions in applications like flood forecasting. In this study, we introduce a new architecture, multi-frequency LSTM (MF-LSTM), designed to use inputs of various temporal frequencies to produce sub-daily (e.g. hourly) predictions at a moderate computational cost.