Articles | Volume 26, issue 12
https://doi.org/10.5194/hess-26-3079-2022
https://doi.org/10.5194/hess-26-3079-2022
Research article
 | 
20 Jun 2022
Research article |  | 20 Jun 2022

Hydrological concept formation inside long short-term memory (LSTM) networks

Thomas Lees, Steven Reece, Frederik Kratzert, Daniel Klotz, Martin Gauch, Jens De Bruijn, Reetik Kumar Sahu, Peter Greve, Louise Slater, and Simon J. Dadson

Related authors

Benchmarking data-driven rainfall–runoff models in Great Britain: a comparison of long short-term memory (LSTM)-based models with four lumped conceptual models
Thomas Lees, Marcus Buechel, Bailey Anderson, Louise Slater, Steven Reece, Gemma Coxon, and Simon J. Dadson
Hydrol. Earth Syst. Sci., 25, 5517–5534, https://doi.org/10.5194/hess-25-5517-2021,https://doi.org/10.5194/hess-25-5517-2021, 2021
Short summary
Nonstationary weather and water extremes: a review of methods for their detection, attribution, and management
Louise J. Slater, Bailey Anderson, Marcus Buechel, Simon Dadson, Shasha Han, Shaun Harrigan, Timo Kelder, Katie Kowal, Thomas Lees, Tom Matthews, Conor Murphy, and Robert L. Wilby
Hydrol. Earth Syst. Sci., 25, 3897–3935, https://doi.org/10.5194/hess-25-3897-2021,https://doi.org/10.5194/hess-25-3897-2021, 2021
Short summary
A pan-African high-resolution drought index dataset
Jian Peng, Simon Dadson, Feyera Hirpa, Ellen Dyer, Thomas Lees, Diego G. Miralles, Sergio M. Vicente-Serrano, and Chris Funk
Earth Syst. Sci. Data, 12, 753–769, https://doi.org/10.5194/essd-12-753-2020,https://doi.org/10.5194/essd-12-753-2020, 2020
Short summary

Related subject area

Subject: Hydrometeorology | Techniques and Approaches: Modelling approaches
High-resolution land surface modelling over Africa: the role of uncertain soil properties in combination with forcing temporal resolution
Bamidele Oloruntoba, Stefan Kollet, Carsten Montzka, Harry Vereecken, and Harrie-Jan Hendricks Franssen
Hydrol. Earth Syst. Sci., 29, 1659–1683, https://doi.org/10.5194/hess-29-1659-2025,https://doi.org/10.5194/hess-29-1659-2025, 2025
Short summary
Investigating the global and regional response of drought to idealized deforestation using multiple global climate models
Yan Li, Bo Huang, Chunping Tan, Xia Zhang, Francesco Cherubini, and Henning W. Rust
Hydrol. Earth Syst. Sci., 29, 1637–1658, https://doi.org/10.5194/hess-29-1637-2025,https://doi.org/10.5194/hess-29-1637-2025, 2025
Short summary
Distribution, trends, and drivers of flash droughts in the United Kingdom
Iván Noguera, Jamie Hannaford, and Maliko Tanguy
Hydrol. Earth Syst. Sci., 29, 1295–1317, https://doi.org/10.5194/hess-29-1295-2025,https://doi.org/10.5194/hess-29-1295-2025, 2025
Short summary
Are dependencies of extreme rainfall on humidity more reliable in convection-permitting climate models?
Geert Lenderink, Nikolina Ban, Erwan Brisson, Ségolène Berthou, Virginia Edith Cortés-Hernández, Elizabeth Kendon, Hayley J. Fowler, and Hylke de Vries
Hydrol. Earth Syst. Sci., 29, 1201–1220, https://doi.org/10.5194/hess-29-1201-2025,https://doi.org/10.5194/hess-29-1201-2025, 2025
Short summary
Leveraging a radar-based disdrometer network to develop a probabilistic precipitation phase model in eastern Canada
Alexis Bédard-Therrien, François Anctil, Julie M. Thériault, Olivier Chalifour, Fanny Payette, Alexandre Vidal, and Daniel F. Nadeau
Hydrol. Earth Syst. Sci., 29, 1135–1158, https://doi.org/10.5194/hess-29-1135-2025,https://doi.org/10.5194/hess-29-1135-2025, 2025
Short summary

Cited articles

Beven, K.: Deep learning, hydrological processes and the uniqueness of place, Hydrol. Process., 34, 3608–3613, https://doi.org/10.1002/hyp.13805, 2020. a
Beven, K. J.: Rainfall-runoff modelling: the primer, John Wiley & Sons, ISBN 978-0-470-71459-1, 2011. a, b
Burnash, R.: The NWS River Forecast System-catchment modeling, in: Computer models of watershed hydrology, Water Resources Publications, 311–366, ISBN 10 1-887201-74-2, 1995. a
Chu, E., Roy, D., and Andreas, J.: Are visual explanations useful? a case study in model-in-the-loop prediction, arXiv preprint: arXiv:2007.12248, 2020. a
Coxon, G., Addor, N., Bloomfield, J., Freer, J., Fry, M., Hannaford, J., Howden, N., Lane, R., Lewis, M., Robinson, E., Wagener, T., and Woods, R.: Catchment attributes and hydro-meteorological timeseries for 671 catchments across Great Britain (CAMELS-GB), UK CEH [data set], https://doi.org/10.5285/8344e4f3-d2ea-44f5-8afa-86d2987543a9, 2020a.  a, b
Download
Short summary
Despite the accuracy of deep learning rainfall-runoff models, we are currently uncertain of what these models have learned. In this study we explore the internals of one deep learning architecture and demonstrate that the model learns about intermediate hydrological stores of soil moisture and snow water, despite never having seen data about these processes during training. Therefore, we find evidence that the deep learning approach learns a physically realistic mapping from inputs to outputs.
Share