The ability of LSTM to model snowmelt versus rainfall generated floods
- 1Norwegian Water Resources and Energy Directorate (Oslo, Norway)
Abstract
One of the most important skills of hydrological models is to simulate timing and magnitude of flood events. Long Short-Term Memory (LSTM) networks are currently among the most successful models for streamflow and flood prediction over large regions. In snow-influenced catchments, which typically comprise a minority in large-scale studies, floods are generated by two distinctly different processes, snowmelt and rainfall. The applicability of hydrological models in such regions is therefore dependent on their ability to represent both types of floods. Nevertheless, flood evaluations of LSTM taking different flood-generating processes into account are currently lacking. This study fills this gap by evaluating the ability of LSTM to model flood peak characteristics separately for snowmelt and rainfall generated floods. The trained LSTM model successfully simulated streamflow time series across the 103 evaluated catchments, with average NSE of 0.84 and average KGE of 0.86 over the unseen evaluation period. LSTM exhibited better performance in the majority of the catchments in terms of flood peak timing and magnitude for both rainfall and snowmelt generated floods when compared to the operational hydrological model in the region (HBV) used as a benchmark. LSTM had a 27 pp higher percentage of correctly simulated peak days for rainfall generated floods as compared to snowmelt generated floods, similar to what was found for HBV (29 pp). LSTM's ability to simulate flood peak magnitudes was similar for the different flood types, with percent errors within 20 % for close to half of the events. False alarm rate and probability of detection results were similar for snowmelt and rainfall generated events, both slightly better than those of mixed events. LSTM outperformed HBV in 70 %–91 % of the catchments in terms of flood peak timing and magnitude for snowmelt and rainfall generated floods. The largest improvements in flood peak magnitudes were found for rainfall generated events, in particular for catchments where HBV exhibited high (>40 % ) absolute errors. Overall, our findings bring confidence that LSTM can improve hydrological services in regions subject to both snowmelt and rainfall generated floods.
Versions
Related
- CMIP6‐Driven CaMa‐Flood Simulations Reveal Anthropogenic Control on Global Flood Extremes via Precipitation and Snowmelt Drivers
- Snowmelt timing constrains green-up but not peak productivity in alpine grasslands of the western Pyrenees
- Higher prevalence of rising climate-driven snowmelt potential of rain-on-snow events in the Eastern Continental US
- IceWatch: a multimodal deep-learning framework for monitoring glacial lake outburst floods, evaluated at Shisper Glacier, Karakoram
- Floods and firms: Vulnerabilities and resilience to natural disasters in Europe
BibTeX
@article{bakke2026ability,
title = {{The ability of LSTM to model snowmelt versus rainfall generated floods}},
author = {Sigrid Jørgensen Bakke and Danielle Marie Barna and Kolbjørn Engeland and Sjur Anders Kolberg and Sunniva Nordeide},
journal = {Hydrology and Earth System Sciences},
year = {2026},
month = oct,
doi = {10.5194/hess-30-6207-2026},
url = {https://hess.copernicus.org/articles/30/6207/2026/},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - The ability of LSTM to model snowmelt versus rainfall generated floods
AU - Sigrid Jørgensen Bakke
AU - Danielle Marie Barna
AU - Kolbjørn Engeland
AU - Sjur Anders Kolberg
AU - Sunniva Nordeide
PY - 2026
DA - 2026/10/06
JO - Hydrology and Earth System Sciences
DO - 10.5194/hess-30-6207-2026
UR - https://hess.copernicus.org/articles/30/6207/2026/
AB - One of the most important skills of hydrological models is to simulate timing and magnitude of flood events. Long Short-Term Memory (LSTM) networks are currently among the most successful models for streamflow and flood prediction over large regions. In snow-influenced catchments, which typically comprise a minority in large-scale studies, floods are generated by two distinctly different processes, snowmelt and rainfall. The applicability of hydrological models in such regions is therefore dependent on their ability to represent both types of floods. Nevertheless, flood evaluations of LSTM taking different flood-generating processes into account are currently lacking. This study fills this gap by evaluating the ability of LSTM to model flood peak characteristics separately for snowmelt and rainfall generated floods. The trained LSTM model successfully simulated streamflow time series across the 103 evaluated catchments, with average NSE of 0.84 and average KGE of 0.86 over the unseen evaluation period. LSTM exhibited better performance in the majority of the catchments in terms of flood peak timing and magnitude for both rainfall and snowmelt generated floods when compared to the operational hydrological model in the region (HBV) used as a benchmark. LSTM had a 27 pp higher percentage of correctly simulated peak days for rainfall generated floods as compared to snowmelt generated floods, similar to what was found for HBV (29 pp). LSTM's ability to simulate flood peak magnitudes was similar for the different flood types, with percent errors within 20 % for close to half of the events. False alarm rate and probability of detection results were similar for snowmelt and rainfall generated events, both slightly better than those of mixed events. LSTM outperformed HBV in 70 %–91 % of the catchments in terms of flood peak timing and magnitude for snowmelt and rainfall generated floods. The largest improvements in flood peak magnitudes were found for rainfall generated events, in particular for catchments where HBV exhibited high (>40 % ) absolute errors. Overall, our findings bring confidence that LSTM can improve hydrological services in regions subject to both snowmelt and rainfall generated floods.
SN - 1607-7938
ER -