Articles | Volume 26, issue 12
https://doi.org/10.5194/hess-26-3079-2022
© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/hess-26-3079-2022
© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Hydrological concept formation inside long short-term memory (LSTM) networks
School of Geography and the Environment, University of Oxford, South Parks Road, Oxford, OX1 3QY, United Kingdom
International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria
Steven Reece
Department of Engineering, University of Oxford, Oxford, United Kingdom
Frederik Kratzert
Google Research, Vienna, Austria
Daniel Klotz
LIT AI Lab & Institute for Machine Learning, Johannes Kepler University Linz, Linz, Austria
Martin Gauch
LIT AI Lab & Institute for Machine Learning, Johannes Kepler University Linz, Linz, Austria
Jens De Bruijn
International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria
Institute for Environmental Studies, VU University, De Boelelaan 1087, 1081 HV Amsterdam, the Netherlands
Reetik Kumar Sahu
International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria
Peter Greve
International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria
Louise Slater
School of Geography and the Environment, University of Oxford, South Parks Road, Oxford, OX1 3QY, United Kingdom
Simon J. Dadson
School of Geography and the Environment, University of Oxford, South Parks Road, Oxford, OX1 3QY, United Kingdom
UK Centre for Ecology and Hydrology, Maclean Building, Crowmarsh Gifford, Wallingford, OX10 8BB, United Kingdom
Related authors
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
Short summary
We used deep learning (DL) models to simulate the amount of water moving through a river channel (discharge) based on the rainfall, temperature and potential evaporation in the previous days. We tested the DL models on catchments across Great Britain finding that the model can accurately simulate hydrological systems across a variety of catchment conditions. Ultimately, the model struggled most in areas where there is chalky bedrock and where human influence on the catchment is large.
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
Short summary
Weather and water extremes have devastating effects each year. One of the principal challenges for society is understanding how extremes are likely to evolve under the influence of changes in climate, land cover, and other human impacts. This paper provides a review of the methods and challenges associated with the detection, attribution, management, and projection of nonstationary weather and water extremes.
Maximillian Van Wyk de Vries, Alexandre Dunant, Amy L. Johnson, Erin L. Harvey, Sihan Li, Katherine Arrell, Jeevan Baniya, Dipak Basnet, Gopi K. Basyal, Nyima Dorjee Bhotia, Simon J. Dadson, Alexander L. Densmore, Tek Bahadur Dong, Mark E. Kincey, Katie Oven, Anuradha Puri, and Nick J. Rosser
Nat. Hazards Earth Syst. Sci., 25, 1937–1942, https://doi.org/10.5194/nhess-25-1937-2025, https://doi.org/10.5194/nhess-25-1937-2025, 2025
Short summary
Short summary
Mapping exposure to landslides is necessary to mitigate risk and reduce vulnerability. In this study, we show that there is a poor correlation between building damage and deaths from landslides, such that the deadliest landslides do not always destroy the most buildings and vice versa. This has important implications for our management of landslide risk.
Simon Moulds, Louise Slater, Louise Arnal, and Andrew W. Wood
Hydrol. Earth Syst. Sci., 29, 2393–2406, https://doi.org/10.5194/hess-29-2393-2025, https://doi.org/10.5194/hess-29-2393-2025, 2025
Short summary
Short summary
Seasonal streamflow forecasts are an important component of flood risk management. Here, we train and test a machine learning model to predict the monthly maximum daily streamflow up to 4 months ahead. We train the model on precipitation and temperature forecasts to produce probabilistic hindcasts for 579 stations across the UK for the period 2004–2016. We show skilful results up to 4 months ahead in many locations, although, in general, the skill declines with increasing lead time.
Emma Ford, Manuela I. Brunner, Hannah Christensen, and Louise Slater
EGUsphere, https://doi.org/10.5194/egusphere-2025-1493, https://doi.org/10.5194/egusphere-2025-1493, 2025
Short summary
Short summary
This study aims to improve prediction and understanding of extreme flood events in UK near-natural catchments. We develop a machine learning framework to assess the contribution of different features to flood magnitude estimation. We find weather patterns are weak predictors and stress the importance of evaluating model performance across and within catchments.
Martin Gauch, Frederik Kratzert, Daniel Klotz, Grey Nearing, Deborah Cohen, and Oren Gilon
EGUsphere, https://doi.org/10.5194/egusphere-2025-1224, https://doi.org/10.5194/egusphere-2025-1224, 2025
Short summary
Short summary
Missing input data are one of the most common challenges when building deep learning hydrological models. We present and analyze different methods that can produce predictions when certain inputs are missing during training or inference. Our proposed strategies provide high accuracy while allowing for more flexible data handling and being robust to outages in operational scenarios.
Eduardo Acuña Espinoza, Frederik Kratzert, Daniel Klotz, Martin Gauch, Manuel Álvarez Chaves, Ralf Loritz, and Uwe Ehret
Hydrol. Earth Syst. Sci., 29, 1749–1758, https://doi.org/10.5194/hess-29-1749-2025, https://doi.org/10.5194/hess-29-1749-2025, 2025
Short summary
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.
Eduardo Acuña Espinoza, Ralf Loritz, Frederik Kratzert, Daniel Klotz, Martin Gauch, Manuel Álvarez Chaves, and Uwe Ehret
Hydrol. Earth Syst. Sci., 29, 1277–1294, https://doi.org/10.5194/hess-29-1277-2025, https://doi.org/10.5194/hess-29-1277-2025, 2025
Short summary
Short summary
Data-driven techniques have shown the potential to outperform process-based models in rainfall–runoff simulations. Hybrid models, combining both approaches, aim to enhance accuracy and maintain interpretability. Expanding the set of test cases to evaluate hybrid models under different conditions, we test their generalization capabilities for extreme hydrological events.
Sanika Baste, Daniel Klotz, Eduardo Acuña Espinoza, Andras Bardossy, and Ralf Loritz
EGUsphere, https://doi.org/10.5194/egusphere-2025-425, https://doi.org/10.5194/egusphere-2025-425, 2025
Short summary
Short summary
This study evaluates the extrapolation performance of Long Short-Term Memory (LSTM) networks in rainfall-runoff modeling, specifically under extreme conditions. The findings reveal that the LSTM cannot predict discharge values beyond a theoretical limit, which is well below the extremity of its training data. This behavior results from the LSTM's gating structures rather than saturation of cell states alone.
Alexandre Dunant, Tom R. Robinson, Alexander L. Densmore, Nick J. Rosser, Ragindra Man Rajbhandari, Mark Kincey, Sihan Li, Prem Raj Awasthi, Max Van Wyk de Vries, Ramesh Guragain, Erin Harvey, and Simon Dadson
Nat. Hazards Earth Syst. Sci., 25, 267–285, https://doi.org/10.5194/nhess-25-267-2025, https://doi.org/10.5194/nhess-25-267-2025, 2025
Short summary
Short summary
Natural hazards like earthquakes often trigger other disasters, such as landslides, creating complex chains of impacts. We developed a risk model using a mathematical approach called hypergraphs to efficiently measure the impact of interconnected hazards. We showed that it can predict broad patterns of damage to buildings and roads from the 2015 Nepal earthquake. The model's efficiency allows it to generate multiple disaster scenarios, even at a national scale, to support preparedness plans.
Daniel Klotz, Peter Miersch, Thiago V. M. do Nascimento, Fabrizio Fenicia, Martin Gauch, and Jakob Zscheischler
Earth Syst. Sci. Data Discuss., https://doi.org/10.5194/essd-2024-450, https://doi.org/10.5194/essd-2024-450, 2025
Preprint under review for ESSD
Short summary
Short summary
Data availability is central to hydrological science. It is the basis for advancing our understanding of hydrological processes, building prediction models, and anticipatory water management. We present a data-driven daily runoff reconstruction product for natural streamflow. We name it EARLS: European aggregated reconstruction for large-sample studies. The reconstructions represent daily simulations of natural streamflow across Europe and cover the period from 1953 to 2020.
Taher Kahil, Safa Baccour, Julian Joseph, Reetik Sahu, Peter Burek, Jia Yi Ng, Samar Asad, Dor Fridman, Jose Albiac, Frank A. Ward, and Yoshihide Wada
Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2024-238, https://doi.org/10.5194/gmd-2024-238, 2024
Revised manuscript under review for GMD
Short summary
Short summary
This study presents the development of the global version of the ECHO hydro-economic model for assessing the economic and environmental performance of water management options. This improved version covers a large number of basins worldwide, includes a detailed representation of irrigated agriculture, and accounts for economic benefits and costs of water use. Results of this study demonstrates the capacity of the model to address emerging water-related research and practical questions.
Claudia Färber, Henning Plessow, Simon Mischel, Frederik Kratzert, Nans Addor, Guy Shalev, and Ulrich Looser
Earth Syst. Sci. Data Discuss., https://doi.org/10.5194/essd-2024-427, https://doi.org/10.5194/essd-2024-427, 2024
Revised manuscript accepted for ESSD
Short summary
Short summary
Large-sample datasets are essential in hydrological science to support modelling studies and advance process understanding. Caravan is a community initiative to create a large-sample hydrology dataset of meteorological forcing data, catchment attributes, and discharge data for catchments around the world. This dataset is a subset of hydrological discharge data and station-based watersheds from the Global Runoff Data Centre (GRDC), which are covered by an open data policy.
Dapeng Feng, Hylke Beck, Jens de Bruijn, Reetik Kumar Sahu, Yusuke Satoh, Yoshihide Wada, Jiangtao Liu, Ming Pan, Kathryn Lawson, and Chaopeng Shen
Geosci. Model Dev., 17, 7181–7198, https://doi.org/10.5194/gmd-17-7181-2024, https://doi.org/10.5194/gmd-17-7181-2024, 2024
Short summary
Short summary
Accurate hydrologic modeling is vital to characterizing water cycle responses to climate change. For the first time at this scale, we use differentiable physics-informed machine learning hydrologic models to simulate rainfall–runoff processes for 3753 basins around the world and compare them with purely data-driven and traditional modeling approaches. This sets a benchmark for hydrologic estimates around the world and builds foundations for improving global hydrologic simulations.
Frederik Kratzert, Martin Gauch, Daniel Klotz, and Grey Nearing
Hydrol. Earth Syst. Sci., 28, 4187–4201, https://doi.org/10.5194/hess-28-4187-2024, https://doi.org/10.5194/hess-28-4187-2024, 2024
Short summary
Short summary
Recently, a special type of neural-network architecture became increasingly popular in hydrology literature. However, in most applications, this model was applied as a one-to-one replacement for hydrology models without adapting or rethinking the experimental setup. In this opinion paper, we show how this is almost always a bad decision and how using these kinds of models requires the use of large-sample hydrology data sets.
Andreas Auer, Martin Gauch, Frederik Kratzert, Grey Nearing, Sepp Hochreiter, and Daniel Klotz
Hydrol. Earth Syst. Sci., 28, 4099–4126, https://doi.org/10.5194/hess-28-4099-2024, https://doi.org/10.5194/hess-28-4099-2024, 2024
Short summary
Short summary
This work examines the impact of temporal and spatial information on the uncertainty estimation of streamflow forecasts. The study emphasizes the importance of data updates and global information for precise uncertainty estimates. We use conformal prediction to show that recent data enhance the estimates, even if only available infrequently. Local data yield reasonable average estimations but fall short for peak-flow events. The use of global data significantly improves these predictions.
Daniel Klotz, Martin Gauch, Frederik Kratzert, Grey Nearing, and Jakob Zscheischler
Hydrol. Earth Syst. Sci., 28, 3665–3673, https://doi.org/10.5194/hess-28-3665-2024, https://doi.org/10.5194/hess-28-3665-2024, 2024
Short summary
Short summary
The evaluation of model performance is essential for hydrological modeling. Using performance criteria requires a deep understanding of their properties. We focus on a counterintuitive aspect of the Nash–Sutcliffe efficiency (NSE) and show that if we divide the data into multiple parts, the overall performance can be higher than all the evaluations of the subsets. Although this follows from the definition of the NSE, the resulting behavior can have unintended consequences in practice.
Rutong Liu, Jiabo Yin, Louise Slater, Shengyu Kang, Yuanhang Yang, Pan Liu, Jiali Guo, Xihui Gu, Xiang Zhang, and Aliaksandr Volchak
Hydrol. Earth Syst. Sci., 28, 3305–3326, https://doi.org/10.5194/hess-28-3305-2024, https://doi.org/10.5194/hess-28-3305-2024, 2024
Short summary
Short summary
Climate change accelerates the water cycle and alters the spatiotemporal distribution of hydrological variables, thus complicating the projection of future streamflow and hydrological droughts. We develop a cascade modeling chain to project future bivariate hydrological drought characteristics over China, using five bias-corrected global climate model outputs under three shared socioeconomic pathways, five hydrological models, and a deep-learning model.
Solomon H. Gebrechorkos, Julian Leyland, Simon J. Dadson, Sagy Cohen, Louise Slater, Michel Wortmann, Philip J. Ashworth, Georgina L. Bennett, Richard Boothroyd, Hannah Cloke, Pauline Delorme, Helen Griffith, Richard Hardy, Laurence Hawker, Stuart McLelland, Jeffrey Neal, Andrew Nicholas, Andrew J. Tatem, Ellie Vahidi, Yinxue Liu, Justin Sheffield, Daniel R. Parsons, and Stephen E. Darby
Hydrol. Earth Syst. Sci., 28, 3099–3118, https://doi.org/10.5194/hess-28-3099-2024, https://doi.org/10.5194/hess-28-3099-2024, 2024
Short summary
Short summary
This study evaluated six high-resolution global precipitation datasets for hydrological modelling. MSWEP and ERA5 showed better performance, but spatial variability was high. The findings highlight the importance of careful dataset selection for river discharge modelling due to the lack of a universally superior dataset. Further improvements in global precipitation data products are needed.
Marcus Buechel, Louise Slater, and Simon Dadson
Hydrol. Earth Syst. Sci., 28, 2081–2105, https://doi.org/10.5194/hess-28-2081-2024, https://doi.org/10.5194/hess-28-2081-2024, 2024
Short summary
Short summary
Afforestation has been proposed internationally, but the hydrological implications of such large increases in the spatial extent of woodland are not fully understood. In this study, we use a land surface model to simulate hydrology across Great Britain with realistic afforestation scenarios and potential climate changes. Countrywide afforestation minimally influences hydrology, when compared to climate change, and reduces low streamflow whilst not lowering the highest flows.
Moctar Dembélé, Mathieu Vrac, Natalie Ceperley, Sander J. Zwart, Josh Larsen, Simon J. Dadson, Grégoire Mariéthoz, and Bettina Schaefli
Proc. IAHS, 385, 121–127, https://doi.org/10.5194/piahs-385-121-2024, https://doi.org/10.5194/piahs-385-121-2024, 2024
Short summary
Short summary
This study assesses the impact of climate change on the timing, seasonality and magnitude of mean annual minimum (MAM) flows and annual maximum flows (AMF) in the Volta River basin (VRB). Several climate change projection data are use to simulate river flow under multiple greenhouse gas emission scenarios. Future projections show that AMF could increase with various magnitude but negligible shift in time across the VRB, while MAM could decrease with up to 14 days of delay in occurrence.
Bailey J. Anderson, Manuela I. Brunner, Louise J. Slater, and Simon J. Dadson
Hydrol. Earth Syst. Sci., 28, 1567–1583, https://doi.org/10.5194/hess-28-1567-2024, https://doi.org/10.5194/hess-28-1567-2024, 2024
Short summary
Short summary
Elasticityrefers to how much the amount of water in a river changes with precipitation. We usually calculate this using average streamflow values; however, the amount of water within rivers is also dependent on stored water sources. Here, we look at how elasticity varies across the streamflow distribution and show that not only do low and high streamflows respond differently to precipitation change, but also these differences vary with water storage availability.
Maximillian Van Wyk de Vries, Sihan Li, Katherine Arrell, Jeevan Baniya, Dipak Basnet, Gopi K. Basyal, Nyima Dorjee Bhotia, Alexander L. Densmore, Tek Bahadur Dong, Alexandre Dunant, Erin L. Harvey, Ganesh K. Jimee, Mark E. Kincey, Katie Oven, Sarmila Paudyal, Dammar Singh Pujara, Anuradha Puri, Ram Shrestha, Nick J. Rosser, and Simon J. Dadson
EGUsphere, https://doi.org/10.5194/egusphere-2024-397, https://doi.org/10.5194/egusphere-2024-397, 2024
Preprint archived
Short summary
Short summary
This study focuses on understanding soil moisture, a key factor for evaluating hillslope stability and landsliding. In Nepal, where landslides are common, we used a computer model to better understand how rapidly soil dries out after the monsoon season. We calibrated the model using field data and found that, by adjusting soil properties, we could predict moisture levels more accurately. This helps understand where landslides might occur, even where direct measurements are not possible.
Jiabo Yin, Louise J. Slater, Abdou Khouakhi, Le Yu, Pan Liu, Fupeng Li, Yadu Pokhrel, and Pierre Gentine
Earth Syst. Sci. Data, 15, 5597–5615, https://doi.org/10.5194/essd-15-5597-2023, https://doi.org/10.5194/essd-15-5597-2023, 2023
Short summary
Short summary
This study presents long-term (i.e., 1940–2022) and high-resolution (i.e., 0.25°) monthly time series of TWS anomalies over the global land surface. The reconstruction is achieved by using a set of machine learning models with a large number of predictors, including climatic and hydrological variables, land use/land cover data, and vegetation indicators (e.g., leaf area index). Our proposed GTWS-MLrec performs overall as well as, or is more reliable than, previous TWS datasets.
Solomon H. Gebrechorkos, Jian Peng, Ellen Dyer, Diego G. Miralles, Sergio M. Vicente-Serrano, Chris Funk, Hylke E. Beck, Dagmawi T. Asfaw, Michael B. Singer, and Simon J. Dadson
Earth Syst. Sci. Data, 15, 5449–5466, https://doi.org/10.5194/essd-15-5449-2023, https://doi.org/10.5194/essd-15-5449-2023, 2023
Short summary
Short summary
Drought is undeniably one of the most intricate and significant natural hazards with far-reaching consequences for the environment, economy, water resources, agriculture, and societies across the globe. In response to this challenge, we have devised high-resolution drought indices. These indices serve as invaluable indicators for assessing shifts in drought patterns and their associated impacts on a global, regional, and local level facilitating the development of tailored adaptation strategies.
Louise J. Slater, Louise Arnal, Marie-Amélie Boucher, Annie Y.-Y. Chang, Simon Moulds, Conor Murphy, Grey Nearing, Guy Shalev, Chaopeng Shen, Linda Speight, Gabriele Villarini, Robert L. Wilby, Andrew Wood, and Massimiliano Zappa
Hydrol. Earth Syst. Sci., 27, 1865–1889, https://doi.org/10.5194/hess-27-1865-2023, https://doi.org/10.5194/hess-27-1865-2023, 2023
Short summary
Short summary
Hybrid forecasting systems combine data-driven methods with physics-based weather and climate models to improve the accuracy of predictions for meteorological and hydroclimatic events such as rainfall, temperature, streamflow, floods, droughts, tropical cyclones, or atmospheric rivers. We review recent developments in hybrid forecasting and outline key challenges and opportunities in the field.
Grey S. Nearing, Daniel Klotz, Jonathan M. Frame, Martin Gauch, Oren Gilon, Frederik Kratzert, Alden Keefe Sampson, Guy Shalev, and Sella Nevo
Hydrol. Earth Syst. Sci., 26, 5493–5513, https://doi.org/10.5194/hess-26-5493-2022, https://doi.org/10.5194/hess-26-5493-2022, 2022
Short summary
Short summary
When designing flood forecasting models, it is necessary to use all available data to achieve the most accurate predictions possible. This manuscript explores two basic ways of ingesting near-real-time streamflow data into machine learning streamflow models. The point we want to make is that when working in the context of machine learning (instead of traditional hydrology models that are based on
bio-geophysics), it is not necessary to use complex statistical methods for injecting sparse data.
Louise J. Slater, Chris Huntingford, Richard F. Pywell, John W. Redhead, and Elizabeth J. Kendon
Earth Syst. Dynam., 13, 1377–1396, https://doi.org/10.5194/esd-13-1377-2022, https://doi.org/10.5194/esd-13-1377-2022, 2022
Short summary
Short summary
This work considers how wheat yields are affected by weather conditions during the three main wheat growth stages in the UK. Impacts are strongest in years with compound weather extremes across multiple growth stages. Future climate projections are beneficial for wheat yields, on average, but indicate a high risk of unseen weather conditions which farmers may struggle to adapt to and mitigate against.
Luca Guillaumot, Mikhail Smilovic, Peter Burek, Jens de Bruijn, Peter Greve, Taher Kahil, and Yoshihide Wada
Geosci. Model Dev., 15, 7099–7120, https://doi.org/10.5194/gmd-15-7099-2022, https://doi.org/10.5194/gmd-15-7099-2022, 2022
Short summary
Short summary
We develop and test the first large-scale hydrological model at regional scale with a very high spatial resolution that includes a water management and groundwater flow model. This study infers the impact of surface and groundwater-based irrigation on groundwater recharge and on evapotranspiration in both irrigated and non-irrigated areas. We argue that water table recorded in boreholes can be used as validation data if water management is well implemented and spatial resolution is ≤ 100 m.
Sella Nevo, Efrat Morin, Adi Gerzi Rosenthal, Asher Metzger, Chen Barshai, Dana Weitzner, Dafi Voloshin, Frederik Kratzert, Gal Elidan, Gideon Dror, Gregory Begelman, Grey Nearing, Guy Shalev, Hila Noga, Ira Shavitt, Liora Yuklea, Moriah Royz, Niv Giladi, Nofar Peled Levi, Ofir Reich, Oren Gilon, Ronnie Maor, Shahar Timnat, Tal Shechter, Vladimir Anisimov, Yotam Gigi, Yuval Levin, Zach Moshe, Zvika Ben-Haim, Avinatan Hassidim, and Yossi Matias
Hydrol. Earth Syst. Sci., 26, 4013–4032, https://doi.org/10.5194/hess-26-4013-2022, https://doi.org/10.5194/hess-26-4013-2022, 2022
Short summary
Short summary
Early flood warnings are one of the most effective tools to save lives and goods. Machine learning (ML) models can improve flood prediction accuracy but their use in operational frameworks is limited. The paper presents a flood warning system, operational in India and Bangladesh, that uses ML models for forecasting river stage and flood inundation maps and discusses the models' performances. In 2021, more than 100 million flood alerts were sent to people near rivers over an area of 470 000 km2.
Juliane Mai, Hongren Shen, Bryan A. Tolson, Étienne Gaborit, Richard Arsenault, James R. Craig, Vincent Fortin, Lauren M. Fry, Martin Gauch, Daniel Klotz, Frederik Kratzert, Nicole O'Brien, Daniel G. Princz, Sinan Rasiya Koya, Tirthankar Roy, Frank Seglenieks, Narayan K. Shrestha, André G. T. Temgoua, Vincent Vionnet, and Jonathan W. Waddell
Hydrol. Earth Syst. Sci., 26, 3537–3572, https://doi.org/10.5194/hess-26-3537-2022, https://doi.org/10.5194/hess-26-3537-2022, 2022
Short summary
Short summary
Model intercomparison studies are carried out to test various models and compare the quality of their outputs over the same domain. In this study, 13 diverse model setups using the same input data are evaluated over the Great Lakes region. Various model outputs – such as streamflow, evaporation, soil moisture, and amount of snow on the ground – are compared using standardized methods and metrics. The basin-wise model outputs and observations are made available through an interactive website.
Jonathan M. Frame, Frederik Kratzert, Daniel Klotz, Martin Gauch, Guy Shalev, Oren Gilon, Logan M. Qualls, Hoshin V. Gupta, and Grey S. Nearing
Hydrol. Earth Syst. Sci., 26, 3377–3392, https://doi.org/10.5194/hess-26-3377-2022, https://doi.org/10.5194/hess-26-3377-2022, 2022
Short summary
Short summary
The most accurate rainfall–runoff predictions are currently based on deep learning. There is a concern among hydrologists that deep learning models may not be reliable in extrapolation or for predicting extreme events. This study tests that hypothesis. The deep learning models remained relatively accurate in predicting extreme events compared with traditional models, even when extreme events were not included in the training set.
Daniel Klotz, Frederik Kratzert, Martin Gauch, Alden Keefe Sampson, Johannes Brandstetter, Günter Klambauer, Sepp Hochreiter, and Grey Nearing
Hydrol. Earth Syst. Sci., 26, 1673–1693, https://doi.org/10.5194/hess-26-1673-2022, https://doi.org/10.5194/hess-26-1673-2022, 2022
Short summary
Short summary
This contribution evaluates distributional runoff predictions from deep-learning-based approaches. We propose a benchmarking setup and establish four strong baselines. The results show that accurate, precise, and reliable uncertainty estimation can be achieved with deep learning.
Moctar Dembélé, Mathieu Vrac, Natalie Ceperley, Sander J. Zwart, Josh Larsen, Simon J. Dadson, Grégoire Mariéthoz, and Bettina Schaefli
Hydrol. Earth Syst. Sci., 26, 1481–1506, https://doi.org/10.5194/hess-26-1481-2022, https://doi.org/10.5194/hess-26-1481-2022, 2022
Short summary
Short summary
Climate change impacts on water resources in the Volta River basin are investigated under various global warming scenarios. Results reveal contrasting changes in future hydrological processes and water availability, depending on greenhouse gas emission scenarios, with implications for floods and drought occurrence over the 21st century. These findings provide insights for the elaboration of regional adaptation and mitigation strategies for climate change.
Manuela I. Brunner and Louise J. Slater
Hydrol. Earth Syst. Sci., 26, 469–482, https://doi.org/10.5194/hess-26-469-2022, https://doi.org/10.5194/hess-26-469-2022, 2022
Short summary
Short summary
Assessing the rarity and magnitude of very extreme flood events occurring less than twice a century is challenging due to the lack of observations of such rare events. Here we develop a new approach, pooling reforecast ensemble members from the European Flood Awareness System to increase the sample size available to estimate the frequency of extreme flood events. We demonstrate that such ensemble pooling produces more robust estimates than observation-based estimates.
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
Short summary
We used deep learning (DL) models to simulate the amount of water moving through a river channel (discharge) based on the rainfall, temperature and potential evaporation in the previous days. We tested the DL models on catchments across Great Britain finding that the model can accurately simulate hydrological systems across a variety of catchment conditions. Ultimately, the model struggled most in areas where there is chalky bedrock and where human influence on the catchment is large.
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
Short summary
Weather and water extremes have devastating effects each year. One of the principal challenges for society is understanding how extremes are likely to evolve under the influence of changes in climate, land cover, and other human impacts. This paper provides a review of the methods and challenges associated with the detection, attribution, management, and projection of nonstationary weather and water extremes.
Frederik Kratzert, Daniel Klotz, Sepp Hochreiter, and Grey S. Nearing
Hydrol. Earth Syst. Sci., 25, 2685–2703, https://doi.org/10.5194/hess-25-2685-2021, https://doi.org/10.5194/hess-25-2685-2021, 2021
Short summary
Short summary
We investigate how deep learning models use different meteorological data sets in the task of (regional) rainfall–runoff modeling. We show that performance can be significantly improved when using different data products as input and further show how the model learns to combine those meteorological input differently across time and space. The results are carefully benchmarked against classical approaches, showing the supremacy of the presented approach.
Martin Gauch, Frederik Kratzert, Daniel Klotz, Grey Nearing, Jimmy Lin, and Sepp Hochreiter
Hydrol. Earth Syst. Sci., 25, 2045–2062, https://doi.org/10.5194/hess-25-2045-2021, https://doi.org/10.5194/hess-25-2045-2021, 2021
Short summary
Short summary
We present multi-timescale Short-Term Memory (MTS-LSTM), a machine learning approach that predicts discharge at multiple timescales within one model. MTS-LSTM is significantly more accurate than the US National Water Model and computationally more efficient than an individual LSTM model per timescale. Further, MTS-LSTM can process different input variables at different timescales, which is important as the lead time of meteorological forecasts often depends on their temporal resolution.
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
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
Coxon, G., Addor, N., Bloomfield, J. P., Freer, J., Fry, M., Hannaford, J., Howden, N. J. K., Lane, R., Lewis, M., Robinson, E. L., Wagener, T., and Woods, R.: CAMELS-GB: hydrometeorological time series and landscape attributes for 671 catchments in Great Britain, Earth Syst. Sci. Data, 12, 2459–2483, https://doi.org/10.5194/essd-12-2459-2020, 2020b. a, b, c
Doshi-Velez, F. and Kim, B.: Towards a rigorous science of interpretable
machine learning, arXiv preprint: arXiv:1702.08608, 2017. a
European Union Digital Strategy: Ethics guidelines for trustworthy AI,
https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai (last access: 7 December 2021), 2019. a
Feng, D., Fang, K., and Shen, C.: Enhancing streamflow forecast and extracting insights using long-short term memory networks with data integration at continental scales, Water Resour. Res., 56, e2019WR026793, https://doi.org/10.1029/2019WR026793, 2020. a
Friedman, J., Hastie, T., and Tibshirani, R.: Regularization paths for
generalized linear models via coordinate descent, J. Stat. Softw., 33, 1, https://doi.org/10.18637/jss.v033.i01, 2010. a, b
Gauch, M. and Lin, J.: A Data Scientist's Guide to Streamflow Prediction,
arxiv: preprint, https://arxiv.org/abs/2006.12975 (last access: 15 July 2021), 2020. a
Gauch, M., Kratzert, F., Klotz, D., Nearing, G., Lin, J., and Hochreiter, S.: Rainfall–runoff prediction at multiple timescales with a single Long Short-Term Memory network, Hydrol. Earth Syst. Sci., 25, 2045–2062, https://doi.org/10.5194/hess-25-2045-2021, 2021. a
Ghorbani, A. and Zou, J.: Neuron shapley: Discovering the responsible neurons, arXiv preprint: arXiv:2002.09815, 2020. a
Gruber, A., Scanlon, T., van der Schalie, R., Wagner, W., and Dorigo, W.: Evolution of the ESA CCI Soil Moisture climate data records and their underlying merging methodology, Earth Syst. Sci. Data, 11, 717–739, https://doi.org/10.5194/essd-11-717-2019, 2019. a
Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A.,
Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D.,,
Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., De Chiara, G., Dahlgren, P., Dee, D., Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer, A., Haimberger, L., Healy, S., Hogan, R. J., and Hólm, E., Janisková, M., Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., de Rosnay, P., Rozum, I., Vamborg, F., Villaume, S., and Thépaut, J.-N.: The ERA5 global reanalysis, Q. J. Roy. Meteorol. Soc., 146, 1999–2049, 2020. a
James, G., Witten, D., Hastie, T., and Tibshirani, R.: An introduction to
statistical learning, in: vol. 112, Springer, ISBN 978-1-4614-7138-7, 2013. a
Karpatne, A., Atluri, G., Faghmous, J. H., Steinbach, M., Banerjee, A.,
Ganguly, A., Shekhar, S., Samatova, N., and Kumar, V.: Theory-guided data
science: A new paradigm for scientific discovery from data, IEEE T. Knowl. Data Eng., 29, 2318–2331, 2017. a
Kingma, D. P. and Ba, J.: Adam: A method for stochastic optimization, arXiv
preprint: arXiv:1412.6980, 2014. a
Kirchner, J. W.: Getting the right answers for the right reasons: Linking
measurements, analyses, and models to advance the science of hydrology, Water
Resour. Res., 42, W03S04, https://doi.org/10.1029/2005WR004362, 2006. a
Kratzert, F.: neuralhydrology/neuralhydrology, Github [code], https://github.com/neuralhydrology/neuralhydrology, last access: 7 June 2022. a
Kratzert, F., Klotz, D., Brenner, C., Schulz, K., and Herrnegger, M.:
Rainfall–runoff modelling using Long Short-Term Memory (LSTM) networks,
Hydrol. Earth Syst. Sci., 22, 6005–6022, https://doi.org/10.5194/hess-22-6005-2018, 2018. a, b, c, d
Kratzert, F., Herrnegger, M., Klotz, D., Hochreiter, S., and Klambauer, G.:
NeuralHydrology – Interpreting LSTMs in Hydrology, Springer International Publishing, Cham, 347–362, https://doi.org/10.1007/978-3-030-28954-6_19, 2019a. a
Kratzert, F., Klotz, D., Herrnegger, M., Sampson, A. K., Hochreiter, S., and
Nearing, G. S.: Toward Improved Predictions in Ungauged Basins: Exploiting
the Power of Machine Learning, Water Resour. Res., 55, 11344–11354, https://doi.org/10.1029/2019WR026065, 2019c. a, b
Kratzert, F., Klotz, D., Shalev, G., Klambauer, G., Hochreiter, S., and
Nearing, G.: Towards learning universal, regional, and local hydrological
behaviors via machine learning applied to large-sample datasets, Hydrol.
Earth Syst. Sci., 23, 5089–5110, https://doi.org/10.5194/hess-23-5089-2019, 2019d. a, b, c, d, e
Lane, R., Coxon, G., Freer, J., and Wagener, T.: Lane_et_al_Benchmark_FUSE_GB, University of Bristo [code], https://doi.org/10.5523/bris.3ma509dlakcf720aw8x82aq4tm, 2020. a
Lees, T.: Hydrological Concept Formation inside Long Short-Term Memory (LSTM) networks, Zenodo [data set], https://doi.org/10.5281/zenodo.5600851, 2021. a
Lees, T.: tommylees112/neuralhydrology, Github [code], https://github.com/tommylees112/neuralhydrology/tree/pixel, last access: 7 June 2022. a
Lees, T., Buechel, M., Anderson, B., Slater, L., Reece, S., Coxon, G., and Dadson, S. J.: Benchmarking data-driven rainfall–runoff models in Great Britain: a comparison of long short-term memory (LSTM)-based models with four lumped conceptual models, Hydrol. Earth Syst. Sci., 25, 5517–5534, https://doi.org/10.5194/hess-25-5517-2021, 2021. a, b, c, d, e, f, g, h
Lipton, Z. C.: The Mythos of Model Interpretability: In machine learning, the
concept of interpretability is both important and slippery, Queue, 16, 31–57, 2018. a
Lundberg, S. M. and Lee, S.-I.: A unified approach to interpreting model
predictions, in: Proceedings of the 31st international conference on neural
information processing systems, 4768–4777, https://proceedings.neurips.cc/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf (last access: 15 July 2021), 2017. a
Ma, K., Feng, D., Lawson, K., Tsai, W.-P., Liang, C., Huang, X., Sharma, A.,
and Shen, C.: Transferring hydrologic data across continents – leveraging US data to improve hydrologic prediction in other countries, Earth and Space
Science Open Archive, p. 28, https://doi.org/10.1002/essoar.10504132.1, 2020. a, b
Manning, C., Widmann, M., Bevacqua, E., Van Loon, A. F., Maraun, D., and Vrac, M.: Soil moisture drought in Europe: a compound event of precipitation and potential evapotranspiration on multiple time scales, J. Hydrometeorol., 19, 1255–1271, 2018. a
Muñoz-Sabater, J., Dutra, E., Agustí-Panareda, A., Albergel, C., Arduini, G., Balsamo, G., Boussetta, S., Choulga, M., Harrigan, S., Hersbach, H., Martens, B., Miralles, D. G., Piles, M., Rodríguez-Fernández, N. J., Zsoter, E., Buontempo, C., and Thépaut, J.-N.: ERA5-Land: a state-of-the-art global reanalysis dataset for land applications, Earth Syst. Sci. Data, 13, 4349–4383, https://doi.org/10.5194/essd-13-4349-2021, 2021. a, b, c, d
Nair, V. and Hinton, G. E.: Rectified linear units improve restricted boltzmann machines, Icml, https://www.cs.toronto.edu/~fritz/absps/reluICML.pdf (last access: 15 July 2021), 2010. a
Nguyen, D.: Comparing automatic and human evaluation of local explanations for text classification, in: Volume 1 (Long Papers), Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 1069–1078, https://aclanthology.org/N18-1097/ (last access: 15 July 2021), 2018. a
Olah, C.: Understanding LSTM Networks – colah's blog,
http://colah.github.io/posts/2015-08-Understanding-LSTMs/ (last access: 15 July 2021), 2016. a
Olah, C., Satyanarayan, A., Johnson, I., Carter, S., Schubert, L., Ye, K., and Mordvintsev, A.: The building blocks of interpretability, Distill, 3, e10, https://doi.org/10.23915/distill.00010, 2018. a
Olah, C., Cammarata, N., Schubert, L., Goh, G., Petrov, M., and Carter, S.:
Zoom in: An introduction to circuits, Distill, 5, e00024-001, https://doi.org/10.23915/distill.00024.001, 2020. a
Reichstein, M., Camps-Valls, G., Stevens, B., Jung, M., Denzler, J.,
Carvalhais, N., and Prabhat, P.: Deep learning and process understanding for data-driven Earth system science, Nature, 566, 195–204, https://doi.org/10.1038/s41586-019-0912-1, 2019.
a
Ribeiro, M. T., Singh, S., and Guestrin, C.: “Why should i trust you?”
Explaining the predictions of any classifier, in: Proceedings of the 22nd ACM
SIGKDD international conference on knowledge discovery and data mining, arxiv: prepint, 1135–1144, https://arxiv.org/abs/1602.04938 (last access: 15 July 2021), 2016. a
Samaniego, L., Thober, S., Kumar, R., Wanders, N., Rakovec, O., Pan, M., Zink, M., Sheffield, J., Wood, E. F., and Marx, A.: Anthropogenic warming
exacerbates European soil moisture droughts, Nat. Clim. Change, 8, 421–426, 2018. a
Shen, C.: A Transdisciplinary Review of Deep Learning Research and Its
Relevance for Water Resources Scientists, Water Resour. Res., 54, 8558–8593, https://doi.org/10.1029/2018WR022643, 2018. a
Sklash, M. G. and Farvolden, R. N.: The role of groundwater in storm runoff,
J. Hydrol., 43, 45–65, 1979. a
Spiegelhalter, D.: Should we trust algorithms?, Harvard Data Science Review, 2, https://doi.org/10.1162/99608f92.cb91a35a, 2020. a
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., and Salakhutdinov,
R.: Dropout: a simple way to prevent neural networks from overfitting, J. Mach. Learn. Res., 15, 1929–1958, 2014. a
UK Statistics Authority: Understanding artificial intelligence ethics and
safety: A guide for the responsible design and implementation of AI systems
in the public sector, SSRN 3403301, 2019. a
Wilby, R., Abrahart, R., and Dawson, C.: Detection of conceptual model
rainfall–runoff processes inside an artificial neural network, Hydrolog. Sci. J., 48, 163–181, 2003. a
Young, P.: Data-based mechanistic modelling of environmental, ecological,
economic and engineering systems, Environ. Model. Softw., 13, 105–122, 1998. a
Young, P.: Top-down and data-based mechanistic modelling of rainfall–flow
dynamics at the catchment scale, Hydrol. Process., 17, 2195–2217, 2003. a
Young, P. C. and Beven, K. J.: Data-based mechanistic modelling and the
rainfall-flow non-linearity, Environmetrics, 5, 335–363, 1994. a
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.
Despite the accuracy of deep learning rainfall-runoff models, we are currently uncertain of what...