Seo et al. (2026) Implementation of a multi-layer snow scheme in the GloSea6 seasonal forecast system: impacts on land–atmosphere interactions and climatological biases

Identification

• Journal: Geoscientific model development
• Year: 2026
• Date: 2026-02-11
• Authors: Eunkyo Seo, Paul A. Dirmeyer, Sunlae Tak
• DOI: 10.5194/gmd-19-1261-2026

Research Groups

• Department of Environmental Atmospheric Sciences, Pukyong National University, Busan, Republic of Korea
• Center for Ocean-Land-Atmosphere Studies, George Mason University, Fairfax, Virginia, USA
• Department of Civil, Urban, Earth, and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan, Republic of Korea

Short Summary

This study implements a multi-layer snow scheme in the GloSea6 seasonal forecast system, demonstrating improved simulation of snow seasonality, land-atmosphere interactions, and reduced climatological biases in temperature and precipitation over the Northern Hemisphere by delaying snowmelt and enhancing evaporative cooling.

Objective

• To assess the impact of a multi-layer snow scheme on climatological biases and land-atmosphere interactions within the GloSea6 seasonal forecast system.
• To evaluate its capability to improve the simulation of snow seasonality, soil moisture, surface air temperature, and precipitation, particularly over Northern Hemisphere mid-latitude regions.

Study Configuration

• Spatial Scale: Global land areas, with a focus on the Northern Hemisphere (40–80° N), particularly mid-latitude snow-affected regions (40–55° N). Model horizontal resolution is 0.56° latitude × 0.83° longitude.
• Temporal Scale: Retrospective forecasts over 24 years (1993–2016) for coupled system comparisons, and 22 years (2001–2022) for offline land surface model experiments. Forecast lead times range from 60 days to 6 months.

Methodology and Data

• Models used:
  • Global Seasonal Forecast System (GloSea) versions 5 (G5 single) and 6 (G6 multi, G6 single).
  • Joint UK Land Environment Simulator (JULES) version 4.7 (GL6.0) and 5.6 (GL8.0).
  • Comparison of single-layer snow scheme (composite snow-soil layer) with a multi-layer snow scheme (up to three dynamic layers).
  • GloSea5 (G5 single) and GloSea6 (G6 multi) are fully coupled atmosphere-land-ocean-sea ice models. An additional experiment (G6 single) implements a single-layer snow scheme in GloSea6 to isolate snow physics impacts.
  • JULES offline experiments (JULES single, JULES multi) are driven by reanalysis and satellite data.
• Data sources:
  • Model Initialization: ERA-interim, JRA-55, ERA5 reanalyses.
  • JULES Offline Forcing: ERA5 reanalysis (atmospheric variables), IMERG version 7 (precipitation).
  • Validation:
    • Surface Air Temperature (2 m): NCEP CPC analysis (daily maximum/minimum, 0.5° resolution).
    • Land Reanalysis: ERA5-Land (soil moisture, heat fluxes, ~0.08° resolution).
    • Snow Water Equivalent (SWE): JRA-3Q reanalysis (0.375° resolution).
    • Surface Soil Moisture (SM): Time-filtered ESA CCI Soil Moisture v08.1 (0.25° resolution).
    • Evaporation, Sensible Heat Flux, Net Radiation: GLEAM version 4.2a (0.1° resolution).
    • Precipitation: MSWEP version 2.8 (0.1° resolution).
  • Other: MODIS satellite products (leaf area index, bare soil albedo), GlobAlbedo surface albedo.

Main Results

• The multi-layer snow scheme (JULES multi, G6 multi) more realistically simulates Northern Hemisphere snow seasonality, delaying the onset of snowmelt by approximately 1 to 2 weeks compared to single-layer schemes.
• This delayed snowmelt postpones springtime evaporation, slows soil moisture depletion, and significantly improves soil moisture memory, particularly in mid-latitude regions.
• During winter, the multi-layer scheme simulates warmer soil temperatures due to enhanced insulation, while during spring snowmelt, it leads to surface cooling due to increased surface albedo and subsequent evaporative cooling.
• Increased soil moisture in G6 multi enhances the partitioning of available energy into latent heat flux, promoting evaporative cooling and mitigating excessive water-limited land-atmosphere coupling.
• This results in reduced near-surface warming biases across the entire diurnal period, especially over mid-latitude regions.
• Model performance in simulating precipitation is improved, with an increase in precipitation occurrence over snow-covered regions, attributed to positive evapotranspiration-precipitation feedback.
• The multi-layer scheme reduces biases in evaporation-precipitation feedback and improves the spatial consistency of both water- and energy-limited land-atmosphere coupling processes.
• The impact of advanced snow physics on land-atmosphere interactions is more pronounced in fully coupled models than in offline land surface model simulations, extending its benefits into the summer season.

Contributions

• Demonstrates the critical value of implementing a multi-layer snowpack scheme in seasonal forecast models for improving the fidelity of temperature and precipitation simulations, and the realism of land-atmosphere interactions, with benefits extending from winter through the subsequent summer.
• Provides a comprehensive comparative analysis of single-layer versus multi-layer snow schemes within both offline land surface models and fully coupled seasonal forecast systems (GloSea5 and GloSea6), effectively isolating the impact of snow physics advancements.
• Quantifies the improvements in key land surface and near-surface atmospheric variables, soil moisture memory, evaporation-precipitation feedback, and land-atmosphere coupling regimes across the Northern Hemisphere.
• Highlights that the observed improvements in climatology due to advanced snow physics are distinct from those arising from other model updates or increased ensemble size, underscoring the fundamental importance of realistic snow representation for subseasonal-to-seasonal forecasting.

Funding

• Korea Meteorological Administration Research and Development program (grant no. RS-2023-00241809)
• National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (grant no. RS-2025-02363044)

Citation

@article{Seo2026Implementation,
  author = {Seo, Eunkyo and Dirmeyer, Paul A. and Tak, Sunlae},
  title = {Implementation of a multi-layer snow scheme in the GloSea6 seasonal forecast system: impacts on land–atmosphere interactions and climatological biases},
  journal = {Geoscientific model development},
  year = {2026},
  doi = {10.5194/gmd-19-1261-2026},
  url = {https://doi.org/10.5194/gmd-19-1261-2026}
}