The ability of LSTM to model snowmelt versus rainfall generated floods

Bakke, Sigrid Jørgensen; Barna, Danielle Marie; Engeland, Kolbjørn; Kolberg, Sjur Anders; Nordeide, Sunniva

doi:10.5194/hess-30-6207-2026

Articles | Volume 30, issue 19

https://doi.org/10.5194/hess-30-6207-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/hess-30-6207-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 30, issue 19

Research article

|

06 Oct 2026

Research article |

| 06 Oct 2026

The ability of LSTM to model snowmelt versus rainfall generated floods

Sigrid Jørgensen Bakke, Danielle Marie Barna, Kolbjørn Engeland, Sjur Anders Kolberg, and Sunniva Nordeide

Data sets

Data for "The ability of LSTM to model snowmelt versus rainfall generated floods" Sigrid Jørgensen Bakke et al. https://doi.org/10.5281/zenodo.22942372

Short summary

Hydrological models need to simulate both rainfall and snowmelt generated floods in regions with snow. We evaluated a deep learning model’s ability to capture timing and magnitude of floods generated by snowmelt and rainfall separately. Timing was better simulated for rainfall than snowmelt generated floods, whereas results for flood peak magnitudes were similar. Compared to an operational model, the deep learning model was better at simulating both flood types in the majority of the catchments.