Richter et al. (2026) Technical note: Literature based approach to estimate future snow

Identification

• Journal: Hydrology and earth system sciences
• Year: 2026
• Date: 2026-02-05
• Authors: Bettina Richter, Christoph Marty
• DOI: 10.5194/hess-30-659-2026

Research Groups

WSL Institute for Snow and Avalanche Research SLF, Davos, Switzerland

Short Summary

This study develops a resource-efficient, literature-based approach to project future snow depths and season lengths in regions lacking downscaled climate projections. By synthesizing existing literature and parameterizing reduction curves based on temperature scenarios and elevation, the method reveals significant declines in snow depth and season length across Swiss Alpine stations under a +2°C warming scenario, particularly at lower elevations.

Objective

• To develop a transferable, literature-based approach for projecting future snow depths and season lengths, applicable to locations with historical snow depth data but lacking high-resolution climate projections.

Study Configuration

• Spatial Scale: European Alps, with specific application to four measurement stations in Switzerland (Weissfluhjoch, Maloja, Saanenmöser, Engelberg) ranging from 1023 meters to 2540 meters above sea level. The approach is trained on data with elevations between 750 meters and 2750 meters.
• Temporal Scale: Reference period for historical snow depth data is 1991–2020 (30 years). Future projections are based on temperature scenarios ranging from +1.1°C to +4.8°C, with a primary application for a +2°C scenario (corresponding to the climate period 2043–2072 for the RCP8.5 scenario). Literature data used for training and validation spans various reference periods (e.g., 1971–2000, 1980–2009) and future projection periods up to 2099.

Methodology and Data

• Models used: A literature-based approach using parameterized reduction curves. A quadratic function (fred(x) = −100 + a −a c2 · (x −b)2) was used to approximate relative future snow depths. Linear regression models were trained to derive reduction parameters (maximum relative future snow depth 'a', shift in peak snow depth timing 'Δb', and relative future season length 'Δc') as functions of temperature scenario and elevation. No complex physically-based snow models or dynamically downscaled regional climate simulations were employed.
• Data sources:
  • Heterogeneous literature on future snow depth and snow water equivalent (SWE) from various studies (e.g., Bülow et al., 2025; Marty et al., 2017; Schmucki et al., 2017; Fiddes et al., 2022; Willibald et al., 2020, 2021; Verfaillie et al., 2018; Kotlarski et al., 2022; Morin et al., 2018), categorized into "Literature-Fit" (for training reduction curves) and "Literature-Validation" datasets.
  • Daily manually measured historical snow depth data from four WSL Institute for Snow and Avalanche Research SLF stations in Switzerland (Weissfluhjoch, Maloja, Saanenmöser, Engelberg) for the winter seasons 1991–2020.
  • Climate reports CH2011 and CH2018 were used to standardize and translate emission scenarios (e.g., RCPs) into corresponding temperature scenarios.

Main Results

• The developed approach successfully parameterized reduction curves using three metrics: (1) maximum relative future snow depth ('a'), (2) shift in the timing of peak snow depth ('Δb'), and (3) relative shortening of the snow season ('Δc').
• Maximum relative future snow depth ('a') decreases with increasing temperature scenarios and increases with elevation, indicating more pronounced snow loss at lower elevations under warmer conditions.
• The shift in peak snow depth timing ('Δb') is predominantly negative, signifying an earlier peak in the snow season, with this shift becoming more pronounced at higher elevations.
• The relative change in season length ('Δc') consistently shows shorter snow seasons in future scenarios, with reductions being more substantial at lower elevations and under higher temperature scenarios.
• Applying the method to four Swiss stations under a +2°C temperature scenario projects significant declines in snow depth and season length. For example, at Weissfluhjoch (2540 m a.s.l.), peak median snow depth decreases from 215 cm to 171 cm, and at Saanenmöser (1390 m a.s.l.), it drops from 64 cm to 36 cm.
• Snow seasons are projected to shorten, with the onset delayed and ablation occurring earlier (e.g., Weissfluhjoch snow season projected to begin approximately two weeks later and end nearly two weeks earlier).
• Validation against published literature data shows that the projections align well with reported ranges and accurately replicate expected elevation-dependent trends, with stronger relative decreases observed for thresholds of deeper snow (e.g., >30 cm) compared to shallower snow (e.g., >5 cm).

Contributions

• Introduces a novel, resource-efficient, and transferable literature-based approach for projecting future snow depth and season length changes in Alpine regions.
• Provides a practical tool for estimating climate change-related snow depth declines in snow-dominated regions that lack highly resolved climate projections.
• Harmonizes heterogeneous literature data by standardizing emission scenarios to temperature scenarios and seasonal periods, improving comparability and communication of results.
• Captures local climatology and site-specific features (e.g., exposure, shading) without requiring computationally complex physically-based snow models or dynamically downscaled regional climate simulations.
• Offers a robust and adaptable framework to support decision-makers in assessing climate impacts and developing adaptation strategies for snow-dependent regions.

Funding

• State Secretariat for Economic Affairs SECO
• Seilbahnen Schweiz
• Speed2Zero (a Joint Initiative co-financed by the ETH Board)

Citation

@article{Richter2026Technical,
  author = {Richter, Bettina and Marty, Christoph},
  title = {Technical note: Literature based approach to estimate future snow},
  journal = {Hydrology and earth system sciences},
  year = {2026},
  doi = {10.5194/hess-30-659-2026},
  url = {https://doi.org/10.5194/hess-30-659-2026}
}