Wu et al. (2026) EcoTWIN 1.0: a fully distributed tracer-aided ecohydrological model tracking water, isotopes, and nutrients

Identification

• Journal: Geoscientific model development
• Year: 2026
• Date: 2026-03-19
• Authors: Songjun Wu, Doerthe Tetzlaff, Yupeng Zheng, C. Soulsby
• DOI: 10.5194/gmd-19-2257-2026

Research Groups

• Department of Ecohydrology and Biogeochemistry, Leibniz Institute of Freshwater Ecology and Inland Fisheries, Berlin, Germany
• Geography Institute and IRI THESys, Humboldt University of Berlin, Berlin, Germany
• School of Environmental Science and Engineering, Southern University of Science and Technology, Shenzhen, China
• Northern Rivers Institute, School of Geosciences, University of Aberdeen, Aberdeen, UK

Short Summary

This paper introduces EcoTWIN 1.0, a fully distributed tracer-aided ecohydrological model that simultaneously tracks water, isotope, and nutrient fluxes. The model demonstrated good performance in reproducing calibrated in-stream targets and uncalibrated internal fluxes across 17 large European catchments, proving its flexibility and transferability for prediction and process inference in diverse terrestrial ecosystems.

Objective

• To develop and validate EcoTWIN, a fully distributed tracer-aided ecohydrological model that leverages stable water isotopes to simultaneously track water, isotope, and nutrient fluxes, thereby enabling more robust water quality modeling and process inference.

Study Configuration

• Spatial Scale: 17 large European catchments, ranging from 10^3 to 10^5 square kilometers, with a spatial resolution of 5 square kilometers.
• Temporal Scale: Daily timesteps from 1980 to 2024 (with the first two years used for spin-up).

Methodology and Data

• Models used: EcoTWIN 1.0 (a fully distributed tracer-aided ecohydrological model implemented in C++), incorporating hydrological, isotopic, and nitrogen modules. Hydrological module includes Green-Ampt for infiltration, degree-day model for snowmelt, and Kinematic wave equation for channel routing. Isotopic module uses the Craig-Gordon model for fractionation. Nitrogen module is modified from mHM-Nitrate and HYPE models.
• Data sources:
  • Calibration: Daily discharge (Global Runoff Data Centre - GRDC), in-stream stable water isotopes (Wateriso and Global Network of Isotopes in Rivers - GNIR), and in-stream nitrate-nitrogen (NO3–N) concentrations (Global Environment Monitoring System - GEMStat).
  • Validation (uncalibrated internal fluxes): ERA5 snow depth (reanalysis), MODIS evapotranspiration (satellite), and GRACE surface water mass anomaly (satellite).
  • Model Setup Inputs: Digital elevation model, flow direction, slope, channel width, channel length, land use type (Winkler et al., 2021), soil type (World Reference Base for Soil Resources - WRB2014), and depth-dependent soil properties (SOILGRIDS).
  • Climatic Forcing: Precipitation, air temperature, potential evapotranspiration, relative humidity (ERA5 and E-OBS reanalysis products).
  • Nitrogen Inputs: Fertilization map (Grizzetti et al., 2021) and nitrate concentration of rainfall (Zhu et al., 2025).

Main Results

• EcoTWIN successfully reproduced observed discharge in all 17 catchments with Kling-Gupta Efficiency (KGE) exceeding 0.5 at most sites.
• Isotopic simulations produced acceptable performances at most sites, with some deviations in Alpine regions attributed to data uncertainties or model simplifications.
• Nitrogen simulations showed comparable performance to existing models, with degraded performance mostly in rivers with low NO3–N concentrations.
• Validation against remote sensing products showed good agreement for uncalibrated internal fluxes:
  • Evapotranspiration (against MODIS) had a coefficient of determination (R2) above 0.5 in most catchments.
  • Snow depth (against ERA5) showed R2 > 0.5 in most regions.
  • Surface water mass anomaly (against GRACE) showed an acceptable fit with R2 above 0.3 in approximately half of the regions, with some degradation in Alpine and coastal areas.
• The model provides spatio-temporally explicit estimations of water ages and travel times, with younger water found in high precipitation/low evapotranspiration regions (e.g., Alpine, north-west coast of Europe) and slower turnover rates in central-west European lowlands.
• A proof of concept using the Damköhler Number demonstrated strong positive correlations between soil water travel time and denitrification rates in agricultural-dominated catchments. Long-term averages of Damköhler numbers often remained below 1 in croplands, indicating that transport processes often dominate over nitrogen removal.

Contributions

• Development of EcoTWIN 1.0, one of the first fully distributed tracer-aided ecohydrological models that simultaneously tracks water, stable isotopes, and nutrient fluxes within an integrated C++ framework, offering significant computational efficiency.
• Demonstrates the model's transferability and robustness across a wide range of geographic and climatic gradients in Europe, from boreal to subtropical climates.
• Provides consistent, spatio-temporally explicit estimations of water ages and travel times, which are crucial for understanding hydrological flow paths and bridging catchment hydrology with water quality processes.
• Introduces a selective disassembly structure, offering alternative conceptualizations for key hydrological processes, enhancing model flexibility and adaptability.
• Presents a novel application of the Damköhler Number in a spatially- and temporally-explicit manner to quantify the interplay between water transport timescales and biogeochemical transformation kinetics, particularly for denitrification.

Funding

• WETSCAPES2.0 project (DFG, German Research Foundation, Project-ID 531801029, TRR 410)
• Leibniz Excellence project ISOSCALE
• "Wasserressourcenpreis 2024" of the Rüdiger Kurt Bode-Foundation
• Leibnitz Association Germany, project Wetland Restoration in Peatlands
• Wallenberg Foundation (WP2023-0001) for International KSLA Guest Professor at SLU

Citation

@article{Wu2026EcoTWIN,
  author = {Wu, Songjun and Tetzlaff, Doerthe and Zheng, Yupeng and Soulsby, C.},
  title = {EcoTWIN 1.0: a fully distributed tracer-aided ecohydrological model tracking water, isotopes, and nutrients},
  journal = {Geoscientific model development},
  year = {2026},
  doi = {10.5194/gmd-19-2257-2026},
  url = {https://doi.org/10.5194/gmd-19-2257-2026}
}