Janzing et al. (2025) Hyper-resolution large-scale hydrological modelling benefits from improved process representation in mountain regions

Identification

• Journal: Hydrology and earth system sciences
• Year: 2025
• Date: 2025-12-08
• Authors: Joren Janzing, Niko Wanders, Marit van Tiel, Barry van Jaarsveld, Dirk N. Karger, Manuela I. Brunner
• DOI: 10.5194/hess-29-7041-2025

Research Groups

• WSL Institute for Snow and Avalanche Research SLF, Davos Dorf, Switzerland
• Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
• Climate Change, Extremes and Natural Hazards in Alpine Regions Research Center CERC, Davos Dorf, Switzerland
• Department of Physical Geography, Utrecht University, Utrecht, the Netherlands
• Laboratory of Hydraulics, Hydrology and Glaciology, ETH Zurich, Zurich, Switzerland
• Swiss Federal Institute for Forest, Snow and Landscape Research WSL, Sion, Switzerland
• Swiss Federal Institute for Forest, Snow and Landscape Research WSL, Birmensdorf, Switzerland

Short Summary

This study enhances the hyper-resolution global hydrological model PCR-GLOBWB 2.0 for mountain regions by improving snow, glacier, and soil process representations. The findings demonstrate that these structural and parametric changes significantly improve discharge simulations, although the quality of meteorological forcing remains a primary determinant of overall model performance.

Objective

• To explore the effect of (1) using different meteorological datasets, (2) improving snow and glacier representations, and (3) changing runoff partitioning in the soil on discharge simulations in the hyper-resolution PCR-GLOBWB 2.0 model, focusing on the larger Alpine domain.

Study Configuration

• Spatial Scale: Regional (larger Alpine domain, latitude: 43–51°, longitude: 3–18°), Hyper-resolution (30 arcsec, approximately 1 km).
• Temporal Scale: 1990–2019 (30 years). Calibration period: 2000–2009. Evaluation periods: 1990–1999 (colder), 2010–2019 (warmer).

Methodology and Data

• Models used: PCR-GLOBWB 2.0 (global hydrological model) with:
  • An extended snow module (seasonally varying degree-day factor, exponential melt function, rain-to-snowfall transition temperature range, lateral snow transport scheme).
  • A new integrated glacier module (static part based on Seibert et al., 2018a; dynamic part using the Δh-parameterization from Huss et al., 2010).
  • Adjusted runoff partitioning scheme in the soil module (reduced top soil layer thickness from 30 cm to 15 cm).
• Data sources:
  • Meteorological Forcing: STANDARD (W5E5 v2.0 downscaled with CHELSA-BIOCLIM+), CHELSA (ERA5 downscaled), CERRA-CHELSA (CERRA-Land precipitation, CERRA temperature downscaled with CHELSA algorithm).
  • Evaluation and Calibration Data:
    • Precipitation: Alpine Gridded Precipitation Dataset (APGD).
    • Discharge: LHDA+ (extended Large-sample hydro-meteorological dataset for the Alps, 2167 stations).
    • Snow Water Equivalent (SWE): OSHD (Switzerland, 1 km), SNOWGRID (Austria, 1 km), CERRA-Land (5.5 km), ERA5-Land (9 km), 1047 Alpine measurement stations.
    • Glaciers: Glacier elevation change maps (GECM, 100 m), World Glacier Monitoring Service (WGMS), Glacier Monitoring Switzerland (GLAMOS), Farinotti et al. (2019) for initial ice thickness, Zekollari et al. (2020) for glacier response time simulations, Randolph Glacier Inventory (RGI) 6.0 for glacier outlines.
    • Soil Moisture: European Space Agency Climate Change Initiative (ESACCI) COMBINED soil moisture data v8.1 (0.25°).
  • Ancillary Data: MERIT Hydro DEM, catchment characteristics (area, reservoirs, glacier fraction, snowfall fraction, potential evapotranspiration).

Main Results

• Meteorological Forcing: CERRA-CHELSA precipitation showed higher correlations with the reference APGD and smaller mean absolute bias (0.4 mm/d) compared to CHELSA (0.5 mm/d) and STANDARD (4 mm/d). However, its impact on discharge simulations was spatially mixed, not consistently outperforming other forcing products across the entire domain.
• Snow Representation:
  • The snow transport scheme improved SWE representation at the highest elevations (2000–3000 m) by reducing unrealistic snow accumulation.
  • Structural changes to the snow module (seasonally varying degree-day factor, exponential melt, rain-to-snowfall transition) improved SWE at higher elevations (KGE increased from 0.85 to 0.89 in Switzerland, and 0.53 to 0.57 in Austria for 2000–3000 m) but led to reduced melt rates at lower elevations in Austria.
  • Calibrating the degree-day factors against Swiss SWE reanalysis resulted in the best SWE performance (KGE of 0.91 in Switzerland, 0.80 in Austria) and improved discharge in most snow-dominated catchments.
• Glacier Representation:
  • The new integrated glacier module successfully captured general glacier behavior, including spatial patterns of elevation changes and mean mass balances, though with biases for individual glaciers.
  • It led to improvements in discharge simulations primarily in highly glacierized catchments.
  • Long-term glacier response simulations showed approximately 40% committed mass loss by 2018, consistent with previous studies.
• Soil Partitioning:
  • Reducing the top soil layer thickness from 30 cm to 15 cm generally increased discharge performance, particularly in rainfall-dominated, unregulated catchments with lower snowfall fractions (<0.3).
  • However, soil moisture representation did not consistently improve, and absolute discharge performance in smaller rainfall-dominated catchments remained limited.
• Overall Model Performance: The new model setup, combining improved snow, glacier, and soil modules, generally increased streamflow simulation performance. Performance was highest in large, natural, snow-covered, and glacierized catchments. Performance decreased in catchments with negative water balance or significant human regulation (e.g., hydropower), indicating limitations in representing human water management.

Contributions

• Provides a comprehensive evaluation of the hyper-resolution PCR-GLOBWB 2.0 model with enhanced process representation (snow, glaciers, soil) in a complex mountain environment (the Alps).
• Quantifies the individual and combined effects of meteorological forcing, improved cryospheric processes, and adjusted soil runoff partitioning on hydrological simulations.
• Develops a new, regionally calibrated setup for PCR-GLOBWB 2.0 that is better suited for mountain hydrology, offering improved spatial detail and bridging the quality gap between national and global snow water equivalent (SWE) products.
• Highlights the critical importance of accurate process representation for hyper-resolution hydrological models and identifies key areas for future development, particularly in human water management and soil heterogeneity.

Funding

• Swiss National Science Foundation (SNSF) through project "Predicting floods and droughts under global change" (grant no. PZ00P2_201818).

Citation

@article{Janzing2025Hyperresolution,
  author = {Janzing, Joren and Wanders, Niko and Tiel, Marit van and Jaarsveld, Barry van and Karger, Dirk N. and Brunner, Manuela I.},
  title = {Hyper-resolution large-scale hydrological modelling benefits from improved process representation in mountain regions},
  journal = {Hydrology and earth system sciences},
  year = {2025},
  doi = {10.5194/hess-29-7041-2025},
  url = {https://doi.org/10.5194/hess-29-7041-2025}
}