Klotz et al. (2026) EARLS: a runoff reconstruction dataset for Europe

Identification

• Journal: Earth system science data
• Year: 2026
• Date: 2026-07-27
• Authors: Daniel Klotz, Peter Miersch, Thiago Victor Medeiros do Nascimento, Fabrizio Fenicia, Corinna Frank, Martin Gauch, Jakob Zscheischler A
• DOI: 10.5194/essd-18-5485-2026

Research Groups

• Department of Compound Environmental Risks, Helmholtz Centre for Environmental Research – UFZ, Leipzig, Germany
• Eawag: Swiss Federal Institute of Aquatic Science and Technology, Dübendorf, Switzerland
• Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
• WSL Institute for Snow and Avalanche Research SLF, Davos Dorf, Switzerland
• Climate Change, Extremes and Natural Hazards in Alpine Regions Research Center CERC, Davos Dorf, Switzerland
• Google Research, Zürich, Switzerland
• Department of Hydro Sciences, TUD Dresden University of Technology, Dresden, Germany
• Center for Scalable Data Analytics and Artificial Intelligence (ScaDS.AI), Dresden/Leipzig, Germany

Short Summary

This study introduces EARLS: a data-driven daily runoff reconstruction dataset for Europe. The dataset provides streamflow reconstructions for over 10,000 basins in Europe, including uncertainty estimates, covering the period from 1953 to 2023.

Objective

• To develop a large-scale hydrological dataset that leverages machine learning to provide accurate and high-resolution data on runoff processes across Europe.
• To create a new generation of datasets that harness the capabilities of deep learning to obtain accurate and high-resolution data on hydrological processes.

Study Configuration

• Spatial Scale: Continental scale, covering the European region (25°–71.5°N × 25°W–45°E).
• Temporal Scale: Daily resolution, spanning from 1953 to 2023.

Methodology and Data

• Models used: Long Short-Term Memory (LSTM) based rainfall-runoff model.
• Data sources:
  • E-OBS dataset for meteorological forcings.
  • HydroATLAS for static basin attributes.
  • EStreams catalog for streamflow observations.

Main Results

• The EARLS LSTM achieves a median Nash-Sutcliffe efficiency (NSE) of 0.66 for the test basins.
• The model performance is related to static and dynamic basin characteristics, with some basins exhibiting low accuracy values due to data quality issues.
• The dataset provides streamflow reconstructions for over 10,000 basins in Europe, including uncertainty estimates.

Contributions

• EARLS represents a new generation of datasets that leverage machine learning to provide accurate and high-resolution data on hydrological processes.
• The study demonstrates the potential of deep learning approaches for large-scale hydrological analyses.

Funding

• This research was funded by [insert funding agency/project codes].

Citation

@article{Klotz2026EARLS,
  author = {Klotz, Daniel and Miersch, Peter and Nascimento, Thiago Victor Medeiros do and Fenicia, Fabrizio and Frank, Corinna and Gauch, Martin and A, Jakob Zscheischler},
  title = {EARLS: a runoff reconstruction dataset for Europe},
  journal = {Earth system science data},
  year = {2026},
  doi = {10.5194/essd-18-5485-2026},
  url = {https://doi.org/10.5194/essd-18-5485-2026}
}