Sarigil et al. (2026) Technical note: Evaluation of snow water equivalent from large-scale land-surface products over Italy

Identification

• Journal: Hydrology and earth system sciences
• Year: 2026
• Date: 2026-11-02
• Authors: Gökhan Sarigil, Mattia Neri, Francesco Avanzi, Elena Toth
• DOI: 10.5194/hess-30-6763-2026

Research Groups

• Department of Civil, Chemical, Environmental, and Materials Engineering, University of Bologna, Bologna, Italy
• CIMA Research Foundation, University Campus of Savona, Savona, Italy

Short Summary

This study evaluates Snow Water Equivalent (SWE) estimates from three large-scale land-surface products (ERA5-Land, CERRA-Land, VHR-REA_IT) across Italy against the IT-SNOW reference dataset. It reveals significant regional differences in product performance, with forcing biases explaining a substantial portion of the observed SWE discrepancies, offering guidance for long-term hydro-climatological assessments.

Objective

• To evaluate the Snow Water Equivalent (SWE) estimates from three large-scale land-surface products (ERA5-Land, CERRA-Land, VHR-REA_IT) across Italy against the high-resolution, observation-constrained IT-SNOW reference dataset, assessing their potential and limitations for long-term hydro-climatological studies and informing product selection in data-scarce regions.

Study Configuration

• Spatial Scale: Italy, focusing on the Alps and Apennines. Data re-gridded to a common 9 km resolution for comparison. Native resolutions: IT-SNOW (~500 m), VHR-REA_IT (2.2 km), CERRA-Land (5.5 km), ERA5-Land (~9 km).
• Temporal Scale: 14 snow-years, from September 2010 to August 2024.

Methodology and Data

• Models used:
  • IT-SNOW: Snow Multidata Mapping and Modeling (S3M) chain (single-layer scheme with hybrid temperature-index and radiation-driven melt).
  • ERA5-Land: CHTESSEL land-surface model (single-layer bulk snow scheme).
  • CERRA-Land: SURFEX v8.1 land-surface model with ISBA-ES multi-layer snow scheme (12 layers).
  • VHR-REA_IT: COSMO-CLM regional atmospheric model and TERRA-ML land-surface model (single bulk snow reservoir) in a fully coupled system.
• Data sources:
  • Reference SWE: IT-SNOW (high-resolution, observation-constrained snow reanalysis for Italy, ~500 m spatial resolution, assimilates in situ ultrasonic snow depth sensors and satellite snow-covered area from Sentinel-2, MODIS/VIIRS, EUMETSAT H SAF H10).
  • Evaluated SWE products:
    • ERA5-Land: Global land-surface reanalysis (~9 km spatial resolution), driven by ERA5 atmospheric forcing.
    • CERRA-Land: European regional land-surface reanalysis (5.5 km spatial resolution), driven by CERRA regional atmospheric reanalysis (which uses ERA5 lateral boundary conditions and gauge-corrected precipitation).
    • VHR-REA_IT: Very High-Resolution Reanalysis for Italy (2.2 km spatial resolution), fully coupled COSMO-CLM atmospheric model and TERRA-ML land-surface model, driven by ERA5 lateral boundary conditions, without regional observational assimilation.
  • Reference meteorological data for forcing evaluation: SCIA (daily gridded precipitation at 10 km spatial resolution, daily minimum and maximum 2 m air temperature at 5 km spatial resolution).
  • Upstream global reanalysis: ERA5 (~31 km spatial resolution), providing atmospheric forcing or lateral boundary conditions for the evaluated products.

Main Results

• No single product consistently performed best across all metrics and Italian regions; forcing biases were identified as a substantial contributor to SWE discrepancies.
• ERA5-Land showed the strongest positive biases in mean annual SWE and snow cover duration (SCD) in the Alps (especially north-western and central sectors), attributed to precipitation overestimation and persistent cold biases. In the Apennines and lowlands, it generally underestimated mean SWE and SCD, consistent with autumn and winter precipitation deficits. Despite these magnitude biases, ERA5-Land achieved the strongest temporal correlations with IT-SNOW daily time series across most of the domain, indicating faithful reproduction of day-to-day variability.
• CERRA-Land reproduced spatial SWE patterns with more detail than ERA5-Land due to its finer resolution. It generally underestimated mean SWE across most subregions, with the main exception of the north-western Alps where positive biases in autumn-winter precipitation and colder conditions led to overestimation. SCD biases were mixed, with moderate Alpine overestimation (strongest in north-western Alps) but negative biases across the Apennines (strongest in northern Apennines) consistent with precipitation deficits. Daily correlation patterns were comparable to ERA5-Land in the Alps but weaker in the Apennines due to more severe precipitation deficits.
• VHR-REA_IT exhibited the poorest overall performance, despite its finest native resolution (2.2 km) and fully coupled architecture. It showed the most spatially widespread underestimation of mean annual SWE and SCD across all subregions (Alps and Apennines). It also had the weakest daily correlations throughout most of Italy, particularly in the Apennines and lower elevations, due to systematic underestimation leading to frequent zero SWE values when snow should be present. These limitations were attributed to the absence of observational assimilation within its regional domain, allowing biases to propagate and be reinforced by land-atmosphere feedbacks.

Contributions

• Provides the first systematic evaluation of Snow Water Equivalent (SWE) estimates from large-scale land-surface products across the entirety of Italy.
• Offers valuable insights for utilizing SWE estimates from large-scale land-surface products in long-term hydro-climatological assessments over Italy.
• Helps characterize the reliability of these products in other hydrologically similar mountainous regions that lack high-quality benchmark estimates.
• Demonstrates that a joint evaluation of SWE products with their driving meteorological fields allows for a more comprehensive interpretation of discrepancies, distinguishing between forcing errors and snow process representation issues.
• Identifies ERA5-Land as the most promising candidate for extending SWE records in Italy due to its high temporal fidelity, despite magnitude biases, provided systematic biases are accounted for.
• Underscores that higher spatial resolution and full land-atmosphere coupling alone do not guarantee more accurate SWE estimates without the incorporation of observational constraints.

Funding

• RETURN Extended Partnership under European Union Next-GenerationEU (National Recovery and Resilience Plan – NRRP, Mission 4, Component 2, Investment 1.3 – D.D. 1243 2/8/2022, PE0000005)
• Project “Stochastic amplification of climate change into floods and droughts change (CO22Water)” – FIS00000846 – CUP J53C23003860001
• Italian Science Fund under FIS1

Citation

@article{Sarigil2026Technical,
  author = {Sarigil, Gökhan and Neri, Mattia and Avanzi, Francesco and Toth, Elena},
  title = {Technical note: Evaluation of snow water equivalent from large-scale land-surface products over Italy},
  journal = {Hydrology and earth system sciences},
  year = {2026},
  doi = {10.5194/hess-30-6763-2026},
  url = {https://doi.org/10.5194/hess-30-6763-2026}
}