Articles | Volume 26, issue 21
https://doi.org/10.5194/hess-26-5493-2022
© 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-26-5493-2022
© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Technical note
| 
04 Nov 2022
Technical note |
| 04 Nov 2022
Technical note: Data assimilation and autoregression for using near-real-time streamflow observations in long short-term memory networks
Grey S. Nearing
CORRESPONDING AUTHOR
Google Research, Mountain View, CA, USA
Daniel Klotz
LIT AI Lab & Institute for Machine Learning, Johannes Kepler University, Linz, Austria
Jonathan M. Frame
Department of Geological Sciences, University of Alabama, Tuscaloosa, AL, USA
Martin Gauch
LIT AI Lab & Institute for Machine Learning, Johannes Kepler University, Linz, Austria
Oren Gilon
Google Research, Tel Aviv, Israel
Frederik Kratzert
Google Research, Vienna, Austria
Alden Keefe Sampson
Upstream Tech, Alameda, CA, USA
Guy Shalev
Google Research, Tel Aviv, Israel
Sella Nevo
Google Research, Tel Aviv, Israel
Related authors
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.
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.
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.
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.
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.
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.
Thomas Lees, Steven Reece, Frederik Kratzert, Daniel Klotz, Martin Gauch, Jens De Bruijn, Reetik Kumar Sahu, Peter Greve, Louise Slater, and Simon J. Dadson
Hydrol. Earth Syst. Sci., 26, 3079–3101, https://doi.org/10.5194/hess-26-3079-2022, https://doi.org/10.5194/hess-26-3079-2022, 2022
Short summary
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.
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.
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.
Frederik Kratzert, Daniel Klotz, Guy Shalev, Günter Klambauer, Sepp Hochreiter, and Grey Nearing
Hydrol. Earth Syst. Sci., 23, 5089–5110, https://doi.org/10.5194/hess-23-5089-2019, https://doi.org/10.5194/hess-23-5089-2019, 2019
Short summary
Short summary
A new approach for regional rainfall–runoff modeling using long short-term memory (LSTM)-based models is presented and benchmarked against a range of well-known hydrological models. The approach significantly outperforms regionally calibrated hydrological models but also basin-wise calibrated models. Furthermore, we propose an adaption of the LSTM that allows us to extract the learned catchment understanding of the model and show that it matches our hydrology expert knowledge.
Frederik Kratzert, Daniel Klotz, Claire Brenner, Karsten Schulz, and Mathew Herrnegger
Hydrol. Earth Syst. Sci., 22, 6005–6022, https://doi.org/10.5194/hess-22-6005-2018, https://doi.org/10.5194/hess-22-6005-2018, 2018
Short summary
Short summary
In this paper, we propose a novel data-driven approach for
rainfall–runoff modelling, using the long short-term memory (LSTM) network, a special type of recurrent neural network. We show in three different experiments that this network is able to learn to predict the discharge purely from meteorological input parameters (such as precipitation or temperature) as accurately as (or better than) the well-established Sacramento Soil Moisture Accounting model, coupled with the Snow-17 snow model.
Karsten Schulz, Reinhard Burgholzer, Daniel Klotz, Johannes Wesemann, and Mathew Herrnegger
Hydrol. Earth Syst. Sci., 22, 2607–2613, https://doi.org/10.5194/hess-22-2607-2018, https://doi.org/10.5194/hess-22-2607-2018, 2018
Short summary
Short summary
The unit hydrograph has been one of the most widely employed modelling techniques to predict rainfall-runoff behaviour of hydrological catchments. We developed a lecture theatre experiment including some student involvement to illustrate the principles behind this modelling technique. The experiment only uses very simple and cheap material including a set of plastic balls (representing rainfall), magnetic stripes (tacking the balls to the white board) and sieves (for ball/water gauging).
Related subject area
Subject: Catchment hydrology | Techniques and Approaches: Modelling approaches
Reconstructing five decades of sediment export from two glacierized high-alpine catchments in Tyrol, Austria, using nonparametric regression
Water and energy budgets over hydrological basins on short and long timescales
Hydrological response to climate change and human activities in the Three-River Source Region
Incorporating experimentally derived streamflow contributions into model parameterization to improve discharge prediction
Machine-learning- and deep-learning-based streamflow prediction in a hilly catchment for future scenarios using CMIP6 GCM data
River hydraulic modeling with ICESat-2 land and water surface elevation
Hydrological modeling using the Soil and Water Assessment Tool in urban and peri-urban environments: the case of Kifisos experimental subbasin (Athens, Greece)
Technical note: How physically based is hydrograph separation by recursive digital filtering?
A comprehensive open-source course for teaching applied hydrological modelling in Central Asia
Impact of distributed meteorological forcing on simulated snow cover and hydrological fluxes over a mid-elevation alpine micro-scale catchment
Technical note: Extending the SWAT model to transport chemicals through tile and groundwater flow
Long-term reconstruction of satellite-based precipitation, soil moisture, and snow water equivalent in China
Disentangling scatter in long-term concentration–discharge relationships: the role of event types
Simulating the hydrological impacts of land use conversion from annual crop to perennial forage in the Canadian Prairies using the Cold Regions Hydrological Modelling platform
How can we benefit from regime information to make more effective use of long short-term memory (LSTM) runoff models?
On the value of satellite remote sensing to reduce uncertainties of regional simulations of the Colorado River
Assessing runoff sensitivity of North American Prairie Pothole Region basins to wetland drainage using a basin classification-based virtual modelling approach
A large-sample investigation into uncertain climate change impacts on high flows across Great Britain
Effects of passive-storage conceptualization on modeling hydrological function and isotope dynamics in the flow system of a cockpit karst landscape
Attribution of climate change and human activities to streamflow variations with a posterior distribution of hydrological simulations
A time-varying distributed unit hydrograph method considering soil moisture
Flood patterns in a catchment with mixed bedrock geology and a hilly landscape: identification of flashy runoff contributions during storm events
A graph neural network (GNN) approach to basin-scale river network learning: the role of physics-based connectivity and data fusion
Improving hydrologic models for predictions and process understanding using neural ODEs
Response of active catchment water storage capacity to a prolonged meteorological drought and asymptotic climate variation
HESS Opinions: Participatory Digital eARth Twin Hydrology systems (DARTHs) for everyone – a blueprint for hydrologists
Development of a national 7-day ensemble streamflow forecasting service for Australia
Future snow changes and their impact on the upstream runoff in Salween
Technical note: Do different projections matter for the Budyko framework?
Revisiting the Hydrological Basis of the Budyko Framework With the Hydrologically Similar Groups Principle
Representation of seasonal land use dynamics in SWAT+ for improved assessment of blue and green water consumption
Large-sample assessment of varying spatial resolution on the streamflow estimates of the wflow_sbm hydrological model
An algorithm for deriving the topology of belowground urban stormwater networks
Assessing the influence of water sampling strategy on the performance of tracer-aided hydrological modeling in a mountainous basin on the Tibetan Plateau
The suitability of differentiable, learnable hydrologic models for ungauged regions and climate change impact assessment
Flood forecasting with machine learning models in an operational framework
Precipitation fate and transport in a Mediterranean catchment through models calibrated on plant and stream water isotope data
High-resolution satellite products improve hydrological modeling in northern Italy
Analysis of high streamflow extremes in climate change studies: how do we calibrate hydrological models?
A conceptual-model-based sediment connectivity assessment for patchy agricultural catchments
The Great Lakes Runoff Intercomparison Project Phase 4: the Great Lakes (GRIP-GL)
Spatial extrapolation of stream thermal peaks using heterogeneous time series at a national scale
Revisiting parameter sensitivities in the variable infiltration capacity model across a hydroclimatic gradient
Deep learning rainfall–runoff predictions of extreme events
Diel streamflow cycles suggest more sensitive snowmelt-driven streamflow to climate change than land surface modeling does
Teaching hydrological modelling: illustrating model structure uncertainty with a ready-to-use computational exercise
Effects of spatial and temporal variability in surface water inputs on streamflow generation and cessation in the rain–snow transition zone
Quantifying multi-year hydrological memory with Catchment Forgetting Curves
On constraining a lumped hydrological model with both piezometry and streamflow: results of a large sample evaluation
Influences of land use changes on the dynamics of water quantity and quality in the German lowland catchment of the Stör
Lena Katharina Schmidt, Till Francke, Peter Martin Grosse, Christoph Mayer, and Axel Bronstert
Hydrol. Earth Syst. Sci., 27, 1841–1863, https://doi.org/10.5194/hess-27-1841-2023, https://doi.org/10.5194/hess-27-1841-2023, 2023
Short summary
Short summary
We present a suitable method to reconstruct sediment export from decadal records of hydroclimatic predictors (discharge, precipitation, temperature) and shorter suspended sediment measurements. This lets us fill the knowledge gap on how sediment export from glacierized high-alpine areas has responded to climate change. We find positive trends in sediment export from the two investigated nested catchments with step-like increases around 1981 which are linked to crucial changes in glacier melt.
Samantha Petch, Bo Dong, Tristan Quaife, Robert P. King, and Keith Haines
Hydrol. Earth Syst. Sci., 27, 1723–1744, https://doi.org/10.5194/hess-27-1723-2023, https://doi.org/10.5194/hess-27-1723-2023, 2023
Short summary
Short summary
Gravitational measurements of water storage from GRACE (Gravity Recovery and Climate Experiment) can improve understanding of the water budget. We produce flux estimates over large river catchments based on observations that close the monthly water budget and ensure consistency with GRACE on short and long timescales. We use energy data to provide additional constraints and balance the long-term energy budget. These flux estimates are important for evaluating climate models.
Ting Su, Chiyuan Miao, Qingyun Duan, Jiaojiao Gou, Xiaoying Guo, and Xi Zhao
Hydrol. Earth Syst. Sci., 27, 1477–1492, https://doi.org/10.5194/hess-27-1477-2023, https://doi.org/10.5194/hess-27-1477-2023, 2023
Short summary
Short summary
The Three-River Source Region (TRSR) plays an extremely important role in water resources security and ecological and environmental protection in China and even all of Southeast Asia. This study used the variable infiltration capacity (VIC) land surface hydrologic model linked with the degree-day factor algorithm to simulate the runoff change in the TRSR. These results will help to guide current and future regulation and management of water resources in the TRSR.
Andreas Hartmann, Jean-Lionel Payeur-Poirier, and Luisa Hopp
Hydrol. Earth Syst. Sci., 27, 1325–1341, https://doi.org/10.5194/hess-27-1325-2023, https://doi.org/10.5194/hess-27-1325-2023, 2023
Short summary
Short summary
We advance our understanding of including information derived from environmental tracers into hydrological modeling. We present a simple approach that integrates streamflow observations and tracer-derived streamflow contributions for model parameter estimation. We consider multiple observed streamflow components and their variation over time to quantify the impact of their inclusion for streamflow prediction at the catchment scale.
Dharmaveer Singh, Manu Vardhan, Rakesh Sahu, Debrupa Chatterjee, Pankaj Chauhan, and Shiyin Liu
Hydrol. Earth Syst. Sci., 27, 1047–1075, https://doi.org/10.5194/hess-27-1047-2023, https://doi.org/10.5194/hess-27-1047-2023, 2023
Short summary
Short summary
This study examines, for the first time, the potential of various machine learning models in streamflow prediction over the Sutlej River basin (rainfall-dominated zone) in western Himalaya during the period 2041–2070 (2050s) and 2071–2100 (2080s) and its relationship to climate variability. The mean ensemble of the model results shows that the mean annual streamflow of the Sutlej River is expected to rise between the 2050s and 2080s by 0.79 to 1.43 % for SSP585 and by 0.87 to 1.10 % for SSP245.
Monica Coppo Frias, Suxia Liu, Xingguo Mo, Karina Nielsen, Heidi Ranndal, Liguang Jiang, Jun Ma, and Peter Bauer-Gottwein
Hydrol. Earth Syst. Sci., 27, 1011–1032, https://doi.org/10.5194/hess-27-1011-2023, https://doi.org/10.5194/hess-27-1011-2023, 2023
Short summary
Short summary
This paper uses remote sensing data from ICESat-2 to calibrate a 1D hydraulic model. With the model, we can make estimations of discharge and water surface elevation, which are important indicators in flooding risk assessment. ICESat-2 data give an added value, thanks to the 0.7 m resolution, which allows the measurement of narrow river streams. In addition, ICESat-2 provides measurements on the river dry portion geometry that can be included in the model.
Evgenia Koltsida, Nikos Mamassis, and Andreas Kallioras
Hydrol. Earth Syst. Sci., 27, 917–931, https://doi.org/10.5194/hess-27-917-2023, https://doi.org/10.5194/hess-27-917-2023, 2023
Short summary
Short summary
Daily and hourly rainfall observations were inputted to a Soil and Water Assessment Tool (SWAT) hydrological model to investigate the impacts of rainfall temporal resolution on a discharge simulation. Results indicated that groundwater flow parameters were more sensitive to daily time intervals, and channel routing parameters were more influential for hourly time intervals. This study suggests that the SWAT model appears to be a reliable tool to predict discharge in a mixed-land-use basin.
Klaus Eckhardt
Hydrol. Earth Syst. Sci., 27, 495–499, https://doi.org/10.5194/hess-27-495-2023, https://doi.org/10.5194/hess-27-495-2023, 2023
Short summary
Short summary
An important hydrological issue is to identify components of streamflow that react to precipitation with different degrees of attenuation and delay. From the multitude of methods that have been developed for this so-called hydrograph separation, a specific, frequently used one is singled out here. It is shown to be derived from plausible physical principles. This increases confidence in its results.
Beatrice Sabine Marti, Aidar Zhumabaev, and Tobias Siegfried
Hydrol. Earth Syst. Sci., 27, 319–330, https://doi.org/10.5194/hess-27-319-2023, https://doi.org/10.5194/hess-27-319-2023, 2023
Short summary
Short summary
Numerical modelling is often used for climate impact studies in water resources management. It is, however, not yet highly accessible to many students of hydrology in Central Asia. One big hurdle for new learners is the preparation of relevant data prior to the actual modelling. We present a robust, open-source workflow and comprehensive teaching material that can be used by teachers and by students for self study.
Aniket Gupta, Alix Reverdy, Jean-Martial Cohard, Basile Hector, Marc Descloitres, Jean-Pierre Vandervaere, Catherine Coulaud, Romain Biron, Lucie Liger, Reed Maxwell, Jean-Gabriel Valay, and Didier Voisin
Hydrol. Earth Syst. Sci., 27, 191–212, https://doi.org/10.5194/hess-27-191-2023, https://doi.org/10.5194/hess-27-191-2023, 2023
Short summary
Short summary
Patchy snow cover during spring impacts mountainous ecosystems on a large range of spatio-temporal scales. A hydrological model simulated such snow patchiness at 10 m resolution. Slope and orientation controls precipitation, radiation, and wind generate differences in snowmelt, subsurface storage, streamflow, and evapotranspiration. The snow patchiness increases the duration of the snowmelt to stream and subsurface storage, which sustains the plants and streamflow later in the summer.
Hendrik Rathjens, Jens Kiesel, Michael Winchell, Jeffrey Arnold, and Robin Sur
Hydrol. Earth Syst. Sci., 27, 159–167, https://doi.org/10.5194/hess-27-159-2023, https://doi.org/10.5194/hess-27-159-2023, 2023
Short summary
Short summary
The SWAT model can simulate the transport of water-soluble chemicals through the landscape but neglects the transport through groundwater or agricultural tile drains. These transport pathways are, however, important to assess the amount of chemicals in streams. We added this capability to the model, which significantly improved the simulation. The representation of all transport pathways in the model enables watershed managers to develop robust strategies for reducing chemicals in streams.
Wencong Yang, Hanbo Yang, Changming Li, Taihua Wang, Ziwei Liu, Qingfang Hu, and Dawen Yang
Hydrol. Earth Syst. Sci., 26, 6427–6441, https://doi.org/10.5194/hess-26-6427-2022, https://doi.org/10.5194/hess-26-6427-2022, 2022
Short summary
Short summary
We produced a daily 0.1° dataset of precipitation, soil moisture, and snow water equivalent in 1981–2017 across China via reconstructions. The dataset used global background data and local on-site data as forcing input and satellite-based data as reconstruction benchmarks. This long-term high-resolution national hydrological dataset is valuable for national investigations of hydrological processes.
Felipe A. Saavedra, Andreas Musolff, Jana von Freyberg, Ralf Merz, Stefano Basso, and Larisa Tarasova
Hydrol. Earth Syst. Sci., 26, 6227–6245, https://doi.org/10.5194/hess-26-6227-2022, https://doi.org/10.5194/hess-26-6227-2022, 2022
Short summary
Short summary
Nitrate contamination of rivers from agricultural sources is a challenge for water quality management. During runoff events, different transport paths within the catchment might be activated, generating a variety of responses in nitrate concentration in stream water. Using nitrate samples from 184 German catchments and a runoff event classification, we show that hydrologic connectivity during runoff events is a key control of nitrate transport from catchments to streams in our study domain.
Marcos R. C. Cordeiro, Kang Liang, Henry F. Wilson, Jason Vanrobaeys, David A. Lobb, Xing Fang, and John W. Pomeroy
Hydrol. Earth Syst. Sci., 26, 5917–5931, https://doi.org/10.5194/hess-26-5917-2022, https://doi.org/10.5194/hess-26-5917-2022, 2022
Short summary
Short summary
This study addresses the issue of increasing interest in the hydrological impacts of converting cropland to perennial forage cover in the Canadian Prairies. By developing customized models using the Cold Regions Hydrological Modelling (CRHM) platform, this long-term (1992–2013) modelling study is expected to provide stakeholders with science-based information regarding the hydrological impacts of land use conversion from annual crop to perennial forage cover in the Canadian Prairies.
Reyhaneh Hashemi, Pierre Brigode, Pierre-André Garambois, and Pierre Javelle
Hydrol. Earth Syst. Sci., 26, 5793–5816, https://doi.org/10.5194/hess-26-5793-2022, https://doi.org/10.5194/hess-26-5793-2022, 2022
Short summary
Short summary
Hydrologists have long dreamed of a tool that could adequately predict runoff in catchments. Data-driven long short-term memory (LSTM) models appear very promising to the hydrology community in this respect. Here, we have sought to benefit from traditional practices in hydrology to improve the effectiveness of LSTM models. We discovered that one LSTM parameter has a hydrologic interpretation and that there is a need to increase the data and to tune two parameters, thereby improving predictions.
Mu Xiao, Giuseppe Mascaro, Zhaocheng Wang, Kristen M. Whitney, and Enrique R. Vivoni
Hydrol. Earth Syst. Sci., 26, 5627–5646, https://doi.org/10.5194/hess-26-5627-2022, https://doi.org/10.5194/hess-26-5627-2022, 2022
Short summary
Short summary
As the major water resource in the southwestern United States, the Colorado River is experiencing decreases in naturalized streamflow and is predicted to face severe challenges under future climate scenarios. Here, we demonstrate the value of Earth observing satellites to improve and build confidence in the spatiotemporal simulations from regional hydrologic models for assessing the sensitivity of the Colorado River to climate change and supporting regional water managers.
Christopher Spence, Zhihua He, Kevin R. Shook, John W. Pomeroy, Colin J. Whitfield, and Jared D. Wolfe
Hydrol. Earth Syst. Sci., 26, 5555–5575, https://doi.org/10.5194/hess-26-5555-2022, https://doi.org/10.5194/hess-26-5555-2022, 2022
Short summary
Short summary
We learnt how streamflow from small creeks could be altered by wetland removal in the Canadian Prairies, where this practice is pervasive. Every creek basin in the region was placed into one of seven groups. We selected one of these groups and used its traits to simulate streamflow. The model worked well enough so that we could trust the results even if we removed the wetlands. Wetland removal did not change low flow amounts very much, but it doubled high flow and tripled average flow.
Rosanna A. Lane, Gemma Coxon, Jim Freer, Jan Seibert, and Thorsten Wagener
Hydrol. Earth Syst. Sci., 26, 5535–5554, https://doi.org/10.5194/hess-26-5535-2022, https://doi.org/10.5194/hess-26-5535-2022, 2022
Short summary
Short summary
This study modelled the impact of climate change on river high flows across Great Britain (GB). Generally, results indicated an increase in the magnitude and frequency of high flows along the west coast of GB by 2050–2075. In contrast, average flows decreased across GB. All flow projections contained large uncertainties; the climate projections were the largest source of uncertainty overall but hydrological modelling uncertainties were considerable in some regions.
Guangxuan Li, Xi Chen, Zhicai Zhang, Lichun Wang, and Chris Soulsby
Hydrol. Earth Syst. Sci., 26, 5515–5534, https://doi.org/10.5194/hess-26-5515-2022, https://doi.org/10.5194/hess-26-5515-2022, 2022
Short summary
Short summary
We developed a coupled flow–tracer model to understand the effects of passive storage on modeling hydrological function and isotope dynamics in a karst flow system. Models with passive storages show improvement in matching isotope dynamics performance, and the improved performance also strongly depends on the number and location of passive storages. Our results also suggested that the solute transport is primarily controlled by advection and hydrodynamic dispersion in the steep hillslope unit.
Xiongpeng Tang, Guobin Fu, Silong Zhang, Chao Gao, Guoqing Wang, Zhenxin Bao, Yanli Liu, Cuishan Liu, and Junliang Jin
Hydrol. Earth Syst. Sci., 26, 5315–5339, https://doi.org/10.5194/hess-26-5315-2022, https://doi.org/10.5194/hess-26-5315-2022, 2022
Short summary
Short summary
In this study, we proposed a new framework that considered the uncertainties of model simulations in quantifying the contribution rate of climate change and human activities to streamflow changes. Then, the Lancang River basin was selected for the case study. The results of quantitative analysis using the new framework showed that the reason for the decrease in the streamflow at Yunjinghong station was mainly human activities.
Bin Yi, Lu Chen, Hansong Zhang, Vijay P. Singh, Ping Jiang, Yizhuo Liu, Hexiang Guo, and Hongya Qiu
Hydrol. Earth Syst. Sci., 26, 5269–5289, https://doi.org/10.5194/hess-26-5269-2022, https://doi.org/10.5194/hess-26-5269-2022, 2022
Short summary
Short summary
An improved GIS-derived distributed unit hydrograph routing method considering time-varying soil moisture was proposed for flow routing. The method considered the changes of time-varying soil moisture and rainfall intensity. The response of underlying surface to the soil moisture content was considered an important factor in this study. The SUH, DUH, TDUH and proposed routing methods (TDUH-MC) were used for flood forecasts, and the simulated results were compared and discussed.
Audrey Douinot, Jean François Iffly, Cyrille Tailliez, Claude Meisch, and Laurent Pfister
Hydrol. Earth Syst. Sci., 26, 5185–5206, https://doi.org/10.5194/hess-26-5185-2022, https://doi.org/10.5194/hess-26-5185-2022, 2022
Short summary
Short summary
The objective of the paper is to highlight the seasonal and singular shift of the transfer time distributions of two catchments (≅10 km2).
Based on 2 years of rainfall and discharge observations, we compare variations in the properties of TTDs with the physiographic characteristics of catchment areas and the eco-hydrological cycle. The paper eventually aims to deduce several factors conducive to particularly rapid and concentrated water transfers, which leads to flash floods.
Alexander Y. Sun, Peishi Jiang, Zong-Liang Yang, Yangxinyu Xie, and Xingyuan Chen
Hydrol. Earth Syst. Sci., 26, 5163–5184, https://doi.org/10.5194/hess-26-5163-2022, https://doi.org/10.5194/hess-26-5163-2022, 2022
Short summary
Short summary
High-resolution river modeling is of great interest to local governments and stakeholders for flood-hazard mitigation. This work presents a physics-guided, machine learning (ML) framework for combining the strengths of high-resolution process-based river network models with a graph-based ML model capable of modeling spatiotemporal processes. Results show that the ML model can approximate the dynamics of the process model with high fidelity, and data fusion further improves the forecasting skill.
Marvin Höge, Andreas Scheidegger, Marco Baity-Jesi, Carlo Albert, and Fabrizio Fenicia
Hydrol. Earth Syst. Sci., 26, 5085–5102, https://doi.org/10.5194/hess-26-5085-2022, https://doi.org/10.5194/hess-26-5085-2022, 2022
Short summary
Short summary
Neural ODEs fuse physics-based models with deep learning: neural networks substitute terms in differential equations that represent the mechanistic structure of the system. The approach combines the flexibility of machine learning with physical constraints for inter- and extrapolation. We demonstrate that neural ODE models achieve state-of-the-art predictive performance while keeping full interpretability of model states and processes in hydrologic modelling over multiple catchments.
Jing Tian, Zhengke Pan, Shenglian Guo, Jiabo Yin, Yanlai Zhou, and Jun Wang
Hydrol. Earth Syst. Sci., 26, 4853–4874, https://doi.org/10.5194/hess-26-4853-2022, https://doi.org/10.5194/hess-26-4853-2022, 2022
Short summary
Short summary
Most of the literature has focused on the runoff response to climate change, while neglecting the impacts of the potential variation in the active catchment water storage capacity (ACWSC) that plays an essential role in the transfer of climate inputs to the catchment runoff. This study aims to systematically identify the response of the ACWSC to a long-term meteorological drought and asymptotic climate change.
Riccardo Rigon, Giuseppe Formetta, Marialaura Bancheri, Niccolò Tubini, Concetta D'Amato, Olaf David, and Christian Massari
Hydrol. Earth Syst. Sci., 26, 4773–4800, https://doi.org/10.5194/hess-26-4773-2022, https://doi.org/10.5194/hess-26-4773-2022, 2022
Short summary
Short summary
The
Digital Earth(DE) metaphor is very useful for both end users and hydrological modelers. We analyse different categories of models, with the view of making them part of a Digital eARth Twin Hydrology system (called DARTH). We also stress the idea that DARTHs are not models in and of themselves, rather they need to be built on an appropriate information technology infrastructure. It is remarked that DARTHs have to, by construction, support the open-science movement and its ideas.
Hapu Arachchige Prasantha Hapuarachchi, Mohammed Abdul Bari, Aynul Kabir, Mohammad Mahadi Hasan, Fitsum Markos Woldemeskel, Nilantha Gamage, Patrick Daniel Sunter, Xiaoyong Sophie Zhang, David Ewen Robertson, James Clement Bennett, and Paul Martinus Feikema
Hydrol. Earth Syst. Sci., 26, 4801–4821, https://doi.org/10.5194/hess-26-4801-2022, https://doi.org/10.5194/hess-26-4801-2022, 2022
Short summary
Short summary
Methodology for developing an operational 7-day ensemble streamflow forecasting service for Australia is presented. The methodology is tested for 100 catchments to learn the characteristics of different NWP rainfall forecasts, the effect of post-processing, and the optimal ensemble size and bootstrapping parameters. Forecasts are generated using NWP rainfall products post-processed by the CHyPP model, the GR4H hydrologic model, and the ERRIS streamflow post-processor inbuilt in the SWIFT package
Chenhao Chai, Lei Wang, Deliang Chen, Jing Zhou, Hu Liu, Jingtian Zhang, Yuanwei Wang, Tao Chen, and Ruishun Liu
Hydrol. Earth Syst. Sci., 26, 4657–4683, https://doi.org/10.5194/hess-26-4657-2022, https://doi.org/10.5194/hess-26-4657-2022, 2022
Short summary
Short summary
This work quantifies future snow changes and their impacts on hydrology in the upper Salween River (USR) under SSP126 and SSP585 using a cryosphere–hydrology model. Future warm–wet climate is not conducive to the development of snow. The rain–snow-dominated pattern of runoff will shift to a rain-dominated pattern after the 2040s under SSP585 but is unchanged under SSP126. The findings improve our understanding of cryosphere–hydrology processes and can assist water resource management in the USR.
Remko C. Nijzink and Stanislaus J. Schymanski
Hydrol. Earth Syst. Sci., 26, 4575–4585, https://doi.org/10.5194/hess-26-4575-2022, https://doi.org/10.5194/hess-26-4575-2022, 2022
Short summary
Short summary
Most catchments plot close to the empirical Budyko curve, which allows for the estimation of the long-term mean annual evaporation and runoff. The Budyko curve can be defined as a function of a wetness index or a dryness index. We found that differences can occur and that there is an uncertainty due to the different formulations.
Yuchan Chen, Xiuzhi Chen, Meimei Xue, Chuanxun Yang, Wei Zheng, Jun Cao, and Wenting Yan
Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2022-290, https://doi.org/10.5194/hess-2022-290, 2022
Revised manuscript accepted for HESS
Short summary
Short summary
This study addresses the quantification and estimation of the watershed characteristic parameter (Pw) in the Budyko framework with the hydrologically similar groups principle. The results show that the Pw is closely related to soil moisture and fractional vegetation cover, and the relationship varies across specific hydrologic similarity groups. The overall satisfactory performance of the Pw estimation model further improves the applicability of the Budyko framework for estimating global runoff.
Anna Msigwa, Celray James Chawanda, Hans C. Komakech, Albert Nkwasa, and Ann van Griensven
Hydrol. Earth Syst. Sci., 26, 4447–4468, https://doi.org/10.5194/hess-26-4447-2022, https://doi.org/10.5194/hess-26-4447-2022, 2022
Short summary
Short summary
Studies using agro-hydrological models, like the Soil and Water Assessment Tool (SWAT), to map evapotranspiration (ET) do not account for cropping seasons. A comparison between the default SWAT+ set-up (with static land use representation) and a dynamic SWAT+ model set-up (with seasonal land use representation) is made by spatial mapping of the ET. The results show that ET with seasonal representation is closer to remote sensing estimates, giving better performance than ET with static land use.
Jerom P. M. Aerts, Rolf W. Hut, Nick C. van de Giesen, Niels Drost, Willem J. van Verseveld, Albrecht H. Weerts, and Pieter Hazenberg
Hydrol. Earth Syst. Sci., 26, 4407–4430, https://doi.org/10.5194/hess-26-4407-2022, https://doi.org/10.5194/hess-26-4407-2022, 2022
Short summary
Short summary
In recent years gridded hydrological modelling moved into the realm of hyper-resolution modelling (<10 km). In this study, we investigate the effect of varying grid-cell sizes for the wflow_sbm hydrological model. We used a large sample of basins from the CAMELS data set to test the effect that varying grid-cell sizes has on the simulation of streamflow at the basin outlet. Results show that there is no single best grid-cell size for modelling streamflow throughout the domain.
Taher Chegini and Hong-Yi Li
Hydrol. Earth Syst. Sci., 26, 4279–4300, https://doi.org/10.5194/hess-26-4279-2022, https://doi.org/10.5194/hess-26-4279-2022, 2022
Short summary
Short summary
Belowground urban stormwater networks (BUSNs) play a critical and irreplaceable role in preventing or mitigating urban floods. However, they are often not available for urban flood modeling at regional or larger scales. We develop a novel algorithm to estimate existing BUSNs using ubiquitously available aboveground data at large scales based on graph theory. The algorithm has been validated in different urban areas; thus, it is well transferable.
Yi Nan, Zhihua He, Fuqiang Tian, Zhongwang Wei, and Lide Tian
Hydrol. Earth Syst. Sci., 26, 4147–4167, https://doi.org/10.5194/hess-26-4147-2022, https://doi.org/10.5194/hess-26-4147-2022, 2022
Short summary
Short summary
Tracer-aided hydrological models are useful tool to reduce uncertainty of hydrological modeling in cold basins, but there is little guidance on the sampling strategy for isotope analysis, which is important for large mountainous basins. This study evaluated the reliance of the tracer-aided modeling performance on the availability of isotope data in the Yarlung Tsangpo river basin, and provides implications for collecting water isotope data for running tracer-aided hydrological models.
Dapeng Feng, Hylke Beck, Kathryn Lawson, and Chaopeng Shen
Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2022-245, https://doi.org/10.5194/hess-2022-245, 2022
Revised manuscript accepted for HESS
Short summary
Short summary
Hybrid models (we call δ models) that embrace the fundamental learning capability of AI but can explain all the physical processes can be powerful. In this paper we assess how they perform when applied in regions not in the training data. δ models rivaled the accuracy of state-of-the-art AI models under the data-dense scenario and even surpassed them for the data-sparse one. They generalize well due to the physical structure. δ models could be ideal candidates for global hydrologic assessments.
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.
Matthias Sprenger, Pilar Llorens, Francesc Gallart, Paolo Benettin, Scott T. Allen, and Jérôme Latron
Hydrol. Earth Syst. Sci., 26, 4093–4107, https://doi.org/10.5194/hess-26-4093-2022, https://doi.org/10.5194/hess-26-4093-2022, 2022
Short summary
Short summary
Our catchment-scale transit time modeling study shows that including stable isotope data on evapotranspiration in addition to the commonly used stream water isotopes helps constrain the model parametrization and reveals that the water taken up by plants has resided longer in the catchment storage than the water leaving the catchment as stream discharge. This finding is important for our understanding of how water is stored and released, which impacts the water availability for plants and humans.
Lorenzo Alfieri, Francesco Avanzi, Fabio Delogu, Simone Gabellani, Giulia Bruno, Lorenzo Campo, Andrea Libertino, Christian Massari, Angelica Tarpanelli, Dominik Rains, Diego G. Miralles, Raphael Quast, Mariette Vreugdenhil, Huan Wu, and Luca Brocca
Hydrol. Earth Syst. Sci., 26, 3921–3939, https://doi.org/10.5194/hess-26-3921-2022, https://doi.org/10.5194/hess-26-3921-2022, 2022
Short summary
Short summary
This work shows advances in high-resolution satellite data for hydrology. We performed hydrological simulations for the Po River basin using various satellite products, including precipitation, evaporation, soil moisture, and snow depth. Evaporation and snow depth improved a simulation based on high-quality ground observations. Interestingly, a model calibration relying on satellite data skillfully reproduces observed discharges, paving the way to satellite-driven hydrological applications.
Bruno Majone, Diego Avesani, Patrick Zulian, Aldo Fiori, and Alberto Bellin
Hydrol. Earth Syst. Sci., 26, 3863–3883, https://doi.org/10.5194/hess-26-3863-2022, https://doi.org/10.5194/hess-26-3863-2022, 2022
Short summary
Short summary
In this work, we introduce a methodology for devising reliable future high streamflow scenarios from climate change simulations. The calibration of a hydrological model is carried out to maximize the probability that the modeled and observed high flow extremes belong to the same statistical population. Application to the Adige River catchment (southeastern Alps, Italy) showed that this procedure produces reliable quantiles of the annual maximum streamflow for use in assessment studies.
Pedro V. G. Batista, Peter Fiener, Simon Scheper, and Christine Alewell
Hydrol. Earth Syst. Sci., 26, 3753–3770, https://doi.org/10.5194/hess-26-3753-2022, https://doi.org/10.5194/hess-26-3753-2022, 2022
Short summary
Short summary
Patchy agricultural landscapes have a large number of small fields, which are separated by linear features such as roads and field borders. When eroded sediments are transported out of the agricultural fields by surface runoff, these features can influence sediment connectivity. By use of measured data and a simulation model, we demonstrate how a dense road network (and its drainage system) facilitates sediment transport from fields to water courses in a patchy Swiss agricultural catchment.
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.
Aurélien Beaufort, Jacob S. Diamond, Eric Sauquet, and Florentina Moatar
Hydrol. Earth Syst. Sci., 26, 3477–3495, https://doi.org/10.5194/hess-26-3477-2022, https://doi.org/10.5194/hess-26-3477-2022, 2022
Short summary
Short summary
We developed one of the largest stream temperature databases to calculate a simple, ecologically relevant metric – the thermal peak – that captures the magnitude of summer thermal extremes. Using statistical models, we extrapolated the thermal peak to nearly every stream in France, finding the hottest thermal peaks along large rivers without forested riparian zones and groundwater inputs. Air temperature was a poor proxy for the thermal peak, highlighting the need to grow monitoring networks.
Ulises M. Sepúlveda, Pablo A. Mendoza, Naoki Mizukami, and Andrew J. Newman
Hydrol. Earth Syst. Sci., 26, 3419–3445, https://doi.org/10.5194/hess-26-3419-2022, https://doi.org/10.5194/hess-26-3419-2022, 2022
Short summary
Short summary
This paper characterizes parameter sensitivities across more than 5500 grid cells for a commonly used macroscale hydrological model, including a suite of eight performance metrics and 43 soil, vegetation and snow parameters. The results show that the model is highly overparameterized and, more importantly, help to provide guidance on the most relevant parameters for specific target processes across diverse climatic types.
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.
Sebastian A. Krogh, Lucia Scaff, James W. Kirchner, Beatrice Gordon, Gary Sterle, and Adrian Harpold
Hydrol. Earth Syst. Sci., 26, 3393–3417, https://doi.org/10.5194/hess-26-3393-2022, https://doi.org/10.5194/hess-26-3393-2022, 2022
Short summary
Short summary
We present a new way to detect snowmelt using daily cycles in streamflow driven by solar radiation. Results show that warmer sites have earlier and more intermittent snowmelt than colder sites, and the timing of early snowmelt events is strongly correlated with the timing of streamflow volume. A space-for-time substitution shows greater sensitivity of streamflow timing to climate change in colder rather than in warmer places, which is then contrasted with land surface simulations.
Wouter J. M. Knoben and Diana Spieler
Hydrol. Earth Syst. Sci., 26, 3299–3314, https://doi.org/10.5194/hess-26-3299-2022, https://doi.org/10.5194/hess-26-3299-2022, 2022
Short summary
Short summary
This paper introduces educational materials that can be used to teach students about model structure uncertainty in hydrological modelling. There are many different hydrological models and differences between these models impact their usefulness in different places. Such models are often used to support decision making about water resources and to perform hydrological science, and it is thus important for students to understand that model choice matters.
Leonie Kiewiet, Ernesto Trujillo, Andrew Hedrick, Scott Havens, Katherine Hale, Mark Seyfried, Stephanie Kampf, and Sarah E. Godsey
Hydrol. Earth Syst. Sci., 26, 2779–2796, https://doi.org/10.5194/hess-26-2779-2022, https://doi.org/10.5194/hess-26-2779-2022, 2022
Short summary
Short summary
Climate change affects precipitation phase, which can propagate into changes in streamflow timing and magnitude. This study examines how variations in rainfall and snowmelt affect discharge. We found that annual discharge and stream cessation depended on the magnitude and timing of rainfall and snowmelt and on the snowpack melt-out date. This highlights the importance of precipitation timing and emphasizes the need for spatiotemporally distributed simulations of snowpack and rainfall dynamics.
Alban de Lavenne, Vazken Andréassian, Louise Crochemore, Göran Lindström, and Berit Arheimer
Hydrol. Earth Syst. Sci., 26, 2715–2732, https://doi.org/10.5194/hess-26-2715-2022, https://doi.org/10.5194/hess-26-2715-2022, 2022
Short summary
Short summary
A watershed remembers the past to some extent, and this memory influences its behavior. This memory is defined by the ability to store past rainfall for several years. By releasing this water into the river or the atmosphere, it tends to forget. We describe how this memory fades over time in France and Sweden. A few watersheds show a multi-year memory. It increases with the influence of groundwater or dry conditions. After 3 or 4 years, they behave independently of the past.
Antoine Pelletier and Vazken Andréassian
Hydrol. Earth Syst. Sci., 26, 2733–2758, https://doi.org/10.5194/hess-26-2733-2022, https://doi.org/10.5194/hess-26-2733-2022, 2022
Short summary
Short summary
A large part of the water cycle takes place underground. In many places, the soil stores water during the wet periods and can release it all year long, which is particularly visible when the river level is low. Modelling tools that are used to simulate and forecast the behaviour of the river struggle to represent this. We improved an existing model to take underground water into account using measurements of the soil water content. Results allow us make recommendations for model users.
Chaogui Lei, Paul D. Wagner, and Nicola Fohrer
Hydrol. Earth Syst. Sci., 26, 2561–2582, https://doi.org/10.5194/hess-26-2561-2022, https://doi.org/10.5194/hess-26-2561-2022, 2022
Short summary
Short summary
We presented an integrated approach to hydrologic modeling and partial least squares regression quantifying land use change impacts on water and nutrient balance over 3 decades. Results highlight that most variations (70 %–80 %) in water quantity and quality variables are explained by changes in land use class-specific areas and landscape metrics. Arable land influences water quantity and quality the most. The study provides insights on water resources management in rural lowland catchments.
Cited articles
Abrahart, R. J. and See, L.: Comparing neural network and autoregressive moving
average techniques for the provision of continuous river flow forecasts in
two contrasting catchments, Hydrol. Proc., 14, 2157–2172, 2000. a
Addor, N., Newman, A. J., Mizukami, N., and Clark, M. P.: The CAMELS data set: catchment attributes and meteorology for large-sample studies, Hydrol. Earth Syst. Sci., 21, 5293–5313, https://doi.org/10.5194/hess-21-5293-2017, 2017.
a, b, c, d
Bannister, R.: A review of operational methods of variational and
ensemble-variational data assimilation, Q. J. Roy. Meteor. Soc., 143, 607–633, 2017. a
Bengio, S., Vinyals, O., Jaitly, N., and Shazeer, N.: Scheduled sampling for
sequence prediction with recurrent neural networks, arXiv [preprint],
https://doi.org/10.48550/arXiv.1506.03099, 2015. a
Cameron, D., Kneale, P., and See, L.: An evaluation of a traditional and a
neural net modelling approach to flood forecasting for an upland catchment,
Hydrol. Proc., 16, 1033–1046, https://doi.org/10.1002/hyp.317, 2002. a
Child, R.: Very deep vaes generalize autoregressive models and can outperform them on
images, arXiv [preprint], https://doi.org/10.48550/arXiv.2011.10650,
2020. a
Chollet, F.: Deep learning with Python, Simon and Schuster, ISBN-13:
9781617296864, 2017. a
De Fauw, J., Dieleman, S., and Simonyan, K.: Hierarchical autoregressive image
models with auxiliary decoders, arXiv [preprint], https://doi.org/10.48550/arXiv.1903.04933, 2019. a
Del Moral, P.: Nonlinear filtering: Interacting particle resolution, Comptes
Rendus de l'Académie des Sciences-Series I-Mathematics, 325, 653–658,
1997. a
Dhariwal, P., Jun, H., Payne, C., Kim, J. W., Radford, A., and Sutskever, I.:
Jukebox: A generative model for music, arXiv [preprint], https://doi.org/10.48550/arXiv.2005.00341, 2020. a
Dong, W., Fong, D. Y. T., Yoon, J.-s., Wan, E. Y. F., Bedford, L. E., Tang, E.
H. M., and Lam, C. L. K.: Generative adversarial networks for imputing
missing data for big data clinical research, BMC Med. Res. Methodol., 21, 1–10, 2021. a
Dosovitskiy, A. and Brox, T.: Inverting visual representations with
convolutional networks, in: Proceedings of the IEEE conference on computer
vision and pattern recognition, 4829–4837, 2016. a
Evensen, G.: The ensemble Kalman filter: Theoretical formulation and practical
implementation, Ocean Dynam., 53, 343–367, 2003. a
Fernandez, B. and Salas, J. D.: Periodic gamma autoregressive processes for
operational hydrology, Water Resour. Res., 22, 1385–1396, 1986. a
Frame, J., Nearing, G., Kratzert, F., and Rahman, M.: Post processing the US national
water model with a long short-term memory network, J. Am. Water Resour. As., https://doi.org/10.31223/osf. io/4xhac, 2020. a
Gaume, E. and Gosset, R.: Over-parameterisation, a major obstacle to the use of artificial neural networks in hydrology?, Hydrol. Earth Syst. Sci., 7, 693–706, https://doi.org/10.5194/hess-7-693-2003, 2003. a
Gers, F. A., Schmidhuber, J., and Cummins, F.: Learning to forget: Continual
prediction with LSTM, Neural Comput., 12, 2451–2471, 2000. a
Graves, A.: Generating sequences with recurrent neural networks, arXiv [preprint], https://doi.org/10.48550/arXiv.1308.0850, 2013. a
Gregor, K., Danihelka, I., Graves, A., Rezende, D., and Wierstra, D.: Draw: A
recurrent neural network for image generation, in: International Conference
on Machine Learning, 1462–1471, 2015. a
Hsu, K.-L., Gupta, H. V., and Sorooshian, S.: Artificial neural network
modeling of the rainfall-runoff process, Water Resour. Res., 31, 2517–2530, 1995. a
Kim, J., Tae, D., and Seok, J.: A survey of missing data imputation using
generative adversarial networks, in: 2020 International Conference on
Artificial Intelligence in Information and Communication (ICAIIC),
IEEE, 454–456, 2020. a
Klotz, D., Kratzert, F., Gauch, M., Keefe Sampson, A., Brandstetter, J., Klambauer, G., Hochreiter, S., and Nearing, G.: Uncertainty estimation with deep learning for rainfall–runoff modeling, Hydrol. Earth Syst. Sci., 26, 1673–1693, https://doi.org/10.5194/hess-26-1673-2022, 2022. a
Kratzert, F.: CAMELS Extended NLDAS Forcing Data, HydroShare [data set],
https://doi.org/10.4211/hs.0a68bfd7ddf642a8be9041d60f40868c, 2019a. a
Kratzert, F.: CAMELS Extended Maurer Forcing Data, HydroShare [data set], https://doi.org/10.4211/hs.17c896843cf940339c3c3496d0c1c077, 2019b. 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
Kratzert, F., Herrnegger, M., Klotz, D., Hochreiter, S., and Klambauer, G.:
Neuralhydrology–interpreting lstms in hydrology, in: Explainable ai:
Interpreting, explaining and visualizing deep learning,
Springer, 347–362, 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,
2019b. a, b, c
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, 2019c. a, b, c, d, e, f, g
Kratzert, F., Gauch, M., Nearing, G., and Klotz, D.: NeuralHydrology – A Python library for Deep Learning research in hydrology, Zenodo [code] https://doi.org/10.5281/zenodo.7063252, 2022a. a
Kratzert, F., Gauch, M., Nearing, G., and Klotz, D.: NeuralHydrology – A Python library for Deep Learning research in hydrology (v.1.3.0), Zenodo [code], https://doi.org/10.5281/zenodo.7063259, 2022b. a
Lamb, A. M., Goyal, A. G. A. P., Zhang, Y., Zhang, S., Courville, A. C., and
Bengio, Y.: Professor forcing: A new algorithm for training recurrent
networks, in: Advances in neural information processing systems,
4601–4609, 2016. a
Mahendran, A. and Vedaldi, A.: Understanding deep image representations by
inverting them, in: Proceedings of the IEEE conference on computer vision and
pattern recognition, 5188–5196, 2015. a
Mai, J., Shen, H., Tolson, B. A., Gaborit, É., Arsenault, R., Craig, J. R., Fortin, V., Fry, L. M., Gauch, M., Klotz, D., Kratzert, F., O'Brien, N., Princz, D. G., Rasiya Koya, S., Roy, T., Seglenieks, F., Shrestha, N. K., Temgoua, A. G. T., Vionnet, V., and Waddell, J. W.: The Great Lakes Runoff Intercomparison Project Phase 4: the Great Lakes (GRIP-GL), Hydrol. Earth Syst. Sci., 26, 3537–3572, https://doi.org/10.5194/hess-26-3537-2022, 2022. a, b
Matalas, N. C.: Mathematical assessment of synthetic hydrology, Water Resour. Res., 3, 937–945, 1967. a
Moshe, Z., Metzger, A., Elidan, G., Kratzert, F., Nevo, S., and El-Yaniv, R.:
Hydronets: Leveraging river structure for hydrologic modeling, arXiv [preprint], https://doi.org/10.48550/arXiv.2007.00595, 2020. a
Nash, J. E. and Sutcliffe, J. V.: River flow forecasting through conceptual
models part I – A discussion of principles, J. Hydrol., 10,
282–290, 1970. a
Nearing, G. S., Gupta, H. V., and Crow, W. T.: Information loss in
approximately Bayesian estimation techniques: A comparison of generative and
discriminative approaches to estimating agricultural productivity, J. Hydrol., 507, 163–173, 2013. a
Nearing, G. S., Kratzert, F., Sampson, A. K., Pelissier, C. S., Klotz, D.,
Frame, J. M., Prieto, C., and Gupta, H. V.: What role does hydrological
science play in the age of machine learning?, Water Resour. Res., 57,
e2020WR028091, https://doi.org/10.1029/2020WR028091, 2020. a
Nevo, S., Morin, E., Gerzi Rosenthal, A., Metzger, A., Barshai, C., Weitzner, D., Voloshin, D., Kratzert, F., Elidan, G., Dror, G., Begelman, G., Nearing, G., Shalev, G., Noga, H., Shavitt, I., Yuklea, L., Royz, M., Giladi, N., Peled Levi, N., Reich, O., Gilon, O., Maor, R., Timnat, S., Shechter, T., Anisimov, V., Gigi, Y., Levin, Y., Moshe, Z., Ben-Haim, Z., Hassidim, A., and Matias, Y.: Flood forecasting with machine learning models in an operational framework, Hydrol. Earth Syst. Sci., 26, 4013–4032, https://doi.org/10.5194/hess-26-4013-2022, 2022. a
Newman, A., Sampson, K., Clark, M. P., Bock, A., Viger, R. J., and Blodgett, D.: A large-sample watershed-scale hydrometeorological dataset for the contiguous USA, UCAR/NCAR [data set], https://doi.org/10.5065/D6MW2F4D, 2014. a
Newman, A. J., Clark, M. P., Sampson, K., Wood, A., Hay, L. E., Bock, A., Viger, R. J., Blodgett, D., Brekke, L., Arnold, J. R., Hopson, T., and Duan, Q.: Development of a large-sample watershed-scale hydrometeorological data set for the contiguous USA: data set characteristics and assessment of regional variability in hydrologic model performance, Hydrol. Earth Syst. Sci., 19, 209–223, https://doi.org/10.5194/hess-19-209-2015, 2015. a
Olah, C., Mordvintsev, A., and Schubert, L.: Feature Visualization, Distill,
https://doi.org/10.23915/distill.00007,
2017. a
Reichle, R. H.: Data assimilation methods in the Earth sciences, Adv. Water Res., 31, 1411–1418, 2008. a
Salinas, D., Flunkert, V., Gasthaus, J., and Januschowski, T.: DeepAR:
Probabilistic forecasting with autoregressive recurrent networks,
Int. J. Forecast., 36, 1181–1191, 2020. a
Szegedy, C., Zaremba, W., Sutskever, I., Bruna, J., Erhan, D., Goodfellow, I.,
and Fergus, R.: Intriguing properties of neural networks, arXiv [preprint], https://doi.org/10.48550/arXiv.1312.6199, 2013.
a
Uria, B., Murray, I., and Larochelle, H.: RNADE: the real-valued neural
autoregressive density-estimator, in: Proceedings of the 26th International
Conference on Neural Information Processing Systems-Volume 2, 2175–2183,
2013. a
van Leeuwen, P. J.: Nonlinear data assimilation in geosciences: an extremely
efficient particle filter, Q. J. Roy. Meteor. Soc., 136, 1991–1999, 2010. a
Van Oord, A., Kalchbrenner, N., and Kavukcuoglu, K.: Pixel recurrent neural
networks, in: International Conference on Machine Learning, PMLR, 1747–1756,
2016. a
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N.,
Kaiser, L., and Polosukhin, I.: Attention is all you need, arXiv [preprint], https://doi.org/10.48550/arXiv.706.03762, 2017. a
Williams, R. J. and Zipser, D.: A learning algorithm for continually running
fully recurrent neural networks, Neural Comput., 1, 270–280, 1989. a
Wunsch, A., Liesch, T., and Broda, S.: Groundwater level forecasting with artificial neural networks: a comparison of long short-term memory (LSTM), convolutional neural networks (CNNs), and non-linear autoregressive networks with exogenous input (NARX), Hydrol. Earth Syst. Sci., 25, 1671–1687, https://doi.org/10.5194/hess-25-1671-2021, 2021. a
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.
When designing flood forecasting models, it is necessary to use all available data to achieve...
