Hossan et al. (2025) Evaluation of wet snow dielectric mixing models for L-band radiometric measurement of liquid water content in Greenland's percolation zone

Identification

• Journal: ˜The œcryosphere
• Year: 2025
• Date: 2025-11-21
• Authors: Alamgir Hossan, Andreas Colliander, Nicole‐Jeanne Schlegel, J. T. Harper, Lauren C. Andrews, Jana Kolassa, Julie Z. Miller, Richard Cullather
• DOI: 10.5194/tc-19-6077-2025

Research Groups

• Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California, United States
• NOAA/OAR Geophysical Fluid Dynamics Laboratory (GFDL), Princeton, New Jersey, United States
• Department of Geosciences, University of Montana, Missoula, Montana, United States
• NASA Global Modeling and Assimilation Office, Goddard Space Flight Center (GSFC), Greenbelt, Maryland, United States
• Science Systems and Applications (SSAI), Berwyn Heights, Maryland, United States
• EarthSAR, LLC, Salt Lake City, Utah, United States
• Cooperative Institute for Research in Environmental Sciences, University of Colorado Boulder, Boulder, Colorado, United States
• Earth System Science Interdisciplinary Center (ESSIC), University of Maryland at College Park, Maryland, United States

Short Summary

This study compares ten microwave dielectric mixing models for estimating liquid water amount (LWA) in wet snow and firn in Greenland's percolation zone using L-band radiometry, finding substantial differences in model performance and LWA retrievals, with power law-based empirical models generally performing better against surface energy and mass balance models.

Objective

• To compare ten commonly used microwave dielectric mixing models for estimating liquid water amount (LWA) in wet snow and firn using L-band radiometry, specifically focusing on the percolation zone of the Greenland Ice Sheet, and evaluate their impact on LWA retrieval accuracy.

Study Configuration

• Spatial Scale: Percolation zone of the Greenland Ice Sheet (GrIS), utilizing SMAP enhanced-resolution data products on an EASE-2 3.125 km grid, and focusing on six Automatic Weather Station (AWS) locations from PROMICE and GC-Net.
• Temporal Scale: The 2023 melt season (June–October for LWA retrieval) with frozen season reference periods (January–March and November–December).

Methodology and Data

• Models used:
  • Dielectric Mixing Models (10 evaluated): Mätzler, Tinga, Debye-like, Hallikainen, Ulaby, Colbeck, Birchak, Sihvola, Looyenga, Tiuri.
  • Radiative Transfer Model: Microwave Emission Model of Layered Snowpacks Version 3 (MEMLS V3).
  • Surface Energy and Mass Balance (SEMB) Models (for validation): Samimi et al. (2021) SEMB model and Glacier Energy and Mass Balance (GEMB) model (within NASA Ice-sheet and Sea-Level System Model (ISSM)).
• Data sources:
  • Satellite: NASA Soil Moisture Active Passive (SMAP) L-band (1.4 GHz) brightness temperature (TB) observations (enhanced-resolution rSIR data products).
  • Observation (in-situ): Automatic Weather Station (AWS) measurements from the Programme for Monitoring of the Greenland Ice Sheet (PROMICE) and Greenland Climate Network (GC-Net), including hourly observations of air temperature, air pressure, radiation fluxes, snow-surface height, wind speeds, and subsurface profiles of temperature, density, and stratigraphy.

Main Results

• The ten dielectric mixing models produce substantially different real and imaginary parts of the dielectric constant for wet snow and firn, leading to significant variations in LWA retrieved from L-band brightness temperature (TB).
• Differences between models are more pronounced for the imaginary part of the dielectric constant and increase with higher liquid water volume fraction (vw), which ranges from 0 % to 6 % in the percolation zone.
• Penetration depth estimates vary significantly among models: for vw of 1 %, 3 %, and 5 %, estimates range between 2.8–12.8 m, 1–4 m, and 0.5–2.3 m, respectively.
• Brightness temperature (TB) sensitivity to LWA changes exponentially, decreasing with increasing LWA, falling below 1 K/mm at less than 50 mm of LWA for both V- and H-polarization.
• The correspondence between LWA retrievals from L-band radiometry and SEMB-derived LWA varied by model and site, with Pearson correlation coefficients ranging from 0.67 to 0.98 and Root Mean Squared Difference (RMSD) values between 5.4 mm and 23.9 mm.
• Power law-based empirical models (Birchak, Sihvola, Looyenga) generally demonstrated better performance, with the Sihvola model showing the best overall agreement (RMSD approximately 11 mm) with SEMB models for the 2023 melt season.
• Satellite retrievals consistently indicate a faster refreezing rate of subsurface liquid water compared to SEMB models, which tend to retain liquid water for longer periods in the post-melt season.

Contributions

• Provides the first comprehensive intercomparison of ten commonly used microwave dielectric mixing models for L-band radiometric estimation of liquid water amount (LWA) in wet snow and firn, specifically in the Greenland Ice Sheet's percolation zone.
• Quantifies the impact of different dielectric mixing formulations on effective permittivity, penetration depth, brightness temperature sensitivity, and LWA retrievals.
• Offers insights into the applicability of models originally derived for higher frequencies (e.g., Hallikainen, Ulaby) to L-band applications.
• Highlights the strengths and weaknesses of various model types (e.g., Maxwell Garnett derivatives, power-law models, empirical models) in the context of L-band LWA retrieval.
• Supports informed selection of dielectric mixing models for improved LWA retrieval accuracy from L-band radiometry.

Funding

• National Aeronautics and Space Administration (NASA)
  • Grant nos. NNH23ZDA001N-MAP
  • Grant nos. NNH22ZDA001N-EUSPI
  • Grant nos. 23-SMAP23-0032

Citation

@article{Hossan2025Evaluation,
  author = {Hossan, Alamgir and Colliander, Andreas and Schlegel, Nicole‐Jeanne and Harper, J. T. and Andrews, Lauren C. and Kolassa, Jana and Miller, Julie Z. and Cullather, Richard},
  title = {Evaluation of wet snow dielectric mixing models for L-band radiometric measurement of liquid water content in Greenland's percolation zone},
  journal = {˜The œcryosphere},
  year = {2025},
  doi = {10.5194/tc-19-6077-2025},
  url = {https://doi.org/10.5194/tc-19-6077-2025}
}