Beguería et al. (2026) Water balance components of the Pyrenees: A 30-year modelling study in a transboundary context

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-03-02
• Authors: Santiago Beguería, Leticia Palazón, Pere Quintana-Seguí, Roger Clavera-Gispert, Anaïs Barella-Ortiz, Marta Domènech, Youen Grusson, Sabine Sauvage, José Miguel Sánchez-Pérez
• DOI: 10.1016/j.ejrh.2026.103195

Research Groups

• Estación Experimental de Aula Dei (EEAD), Consejo Superior de Investigaciones Científicas (CSIC), Zaragoza, Spain
• Observatori de l’Ebre (Universitat Ramon Llull-CSIC), Roquetes, Spain
• Centre de Recherche sur la Biodiversité et l′Environnement (CRBE), Université de Toulouse, CNRS, IRD, Toulouse INP, Toulouse, France

Short Summary

This study reconstructed the regional water balance for the Pyrenees over the 1981–2010 historical baseline using two contrasting hydrological models, SASER and SWAT. Results reveal strong hydroclimatic gradients and highlight evapotranspiration, recharge, and snowmelt timing as key sources of structural uncertainty, establishing the first integrated, transboundary hydrological baseline for the region.

Objective

• To characterize the main water balance components across the Pyrenean region.
• To identify spatial and temporal patterns in hydrological behavior.
• To quantify and interpret the main sources of epistemic uncertainty associated with model structure and process representation.

Study Configuration

• Spatial Scale: The Pyrenees mountain range (transboundary, spanning France, Spain, and Andorra), covering approximately 400 kilometers from the Atlantic to the Mediterranean. Hydrological simulations were performed on a 2.5 km × 2.5 km grid.
• Temporal Scale: Historical baseline period of 1981–2010 (30 hydrological years, October 1980 to September 2010). A five-year warm-up period (1975–1979) was used for model initialization.

Methodology and Data

• Models used:
  • SASER (SAfran–Surfex–Eaudyssée–Rapid): A fully distributed, physically based land surface and routing model. It integrates SURFEX (with ISBA-DIF scheme for soil-vegetation-atmosphere interactions and Explicit Snow (ES) three-layer snow scheme), RAPID (Muskingum-based routing module), and EauDyssée (coupling framework). It was run without calibration to observed discharge.
  • SWAT (Soil and Water Assessment Tool): A semi-distributed, basin-scale, process-based conceptual model. Applied to 20 major river basins, delineated into sub-basins and Hydrological Response Units (HRUs). Calibrated against daily streamflow observations from 87 near-natural gauging stations (1986–2005) and validated (2006–2013).
• Data sources:
  • Meteorological Forcing: PIRAGUAatmosanalysis, a high-resolution (2.5 km × 2.5 km) gridded daily dataset of meteorological variables (precipitation, air temperature, wind speed, relative humidity) derived from the SAFRAN system, harmonizing French and Spanish domains.
  • Evaluation Data:
    • MODIS global actual evapotranspiration (ET) estimates (MOD16) at monthly scale (2000–2010).
    • MODIS snow cover data (MOD10A1) at daily 500 m resolution (2000–2010).
    • Streamflow observations from 87 near-natural gauging stations.

Main Results

• Hydroclimatic Gradients: Precipitation ranges from over 2000 mm/yr in Atlantic headwaters to below 1000 mm/yr in southern and eastern foothills. Mean annual temperature ranges from approximately 4 °C in high elevations to over 14 °C in foothills.
• Snow Dynamics: Snowfall and snowmelt dominate high central and western ranges, sustaining summer flows. Mean annual snowmelt exceeds 600 mm/yr in high ranges. The relative contribution of snowmelt to total water yield shows a small but statistically significant upward trend (+0.5 % per decade), driven by declining rainfall-driven runoff rather than enhanced melt.
• Evapotranspiration (ET): The largest water loss flux, averaging 400–800 mm/yr, with a northwest–southeast gradient. Rising potential evapotranspiration (PET) and declining actual-to-potential evapotranspiration (ET/PET) ratios indicate a shift towards increasingly water-limited conditions and emerging drought stress. SASER simulates higher ET (mean ~720 mm/yr) than SWAT (mean ~580 mm/yr).
• Groundwater Recharge: Modest (< 150 mm/yr in wetter basins, < 50 mm/yr in drier ones, as estimated by SWAT) and shows a significant downward trend (–0.5 to –1.0 mm/yr), suggesting reduced groundwater resilience. SASER does not explicitly model deep aquifer recharge.
• Water Yield and Streamflow: Water yield peaks near 1500 mm/yr in the humid northwest and falls below 200 mm/yr in the arid southeast, showing a slight downward trend (–1 to –2 mm/yr). SWAT systematically simulates higher water yields (mean ~580 mm/yr) than SASER (mean ~400 mm/yr). Streamflow also exhibits a slight downward tendency (SWAT: –167 Mm³/yr; SASER: –32.2 Mm³/yr).
• Model Divergences and Uncertainty: Despite consistent large-scale patterns, significant differences in magnitude and timing of ET, snowmelt, and recharge exist between models, representing epistemic uncertainty due to structural differences (e.g., SASER's energy balance snow scheme vs. SWAT's degree-day approach; SASER's resistance-based ET vs. SWAT's PET-capped ET).
• Evaluation: SWAT achieved robust streamflow performance (median KGE of 0.75 during calibration), while SASER, uncalibrated, showed lower efficiencies (median KGE ≈0.45). Both models showed weak agreement with MODIS ET and snow products in magnitude and seasonal timing, indicating that discharge calibration does not guarantee internal flux realism.

Contributions

• Establishes the first integrated, transboundary hydrological baseline for the Pyrenees over the 1981–2010 period, providing a crucial reference for diagnosing recent changes and evaluating future scenarios.
• Quantifies and interprets epistemic uncertainty in mountain hydrology by comparing two structurally contrasting models (SASER and SWAT) under identical meteorological forcing.
• Provides an openly available dataset (PIRAGUAhydroanalysis) and analysis code, fostering reproducibility and transparency in transboundary water resource assessments.
• Offers a foundation for adaptive and cooperative water management strategies across France, Spain, and Andorra, and a methodological template for other data-scarce mountain regions globally.

Funding

• Interreg V Spain-France-Andorre Programme (POCTEFA 2014–2020)
• European Regional Development Fund (ERDF)

Citation

@article{Beguería2026Water,
  author = {Beguería, Santiago and Palazón, Leticia and Quintana-Seguí, Pere and Clavera-Gispert, Roger and Barella-Ortiz, Anaïs and Domènech, Marta and Grusson, Youen and Sauvage, Sabine and Sánchez-Pérez, José Miguel},
  title = {Water balance components of the Pyrenees: A 30-year modelling study in a transboundary context},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103195},
  url = {https://doi.org/10.1016/j.ejrh.2026.103195}
}