Sign in Subscribe
Machine learning

Breakthrough in Snow Parameter Retrieval Algorithm for Global Climate Modeling

Strategic Insights

— 1 min read

Research conducted by a team of scientists from Stevens Institute of Technology has made a significant breakthrough in the development of a new snow parameter retrieval (SPR) algorithm for the Global Change Observation Mission-Climate/Second Generation Global Imager (GCOM-C/SGLI) instrument. The algorithm, which combines accurate radiative transfer model (RTM) simulations and Scientific Machine Learning (SciML) methods, provides precise retrievals of snow grain size, impurity concentration, and albedo. The research has been published in the journal Frontiers in Environmental Science, and the findings have been validated over the Greenland ice sheet.

Key Takeaways:

  • The new SPR algorithm combines RTM simulations with SciML methods, specifically Multi-Layer Neural-Network (MLNN) techniques, to provide accurate retrievals of snow parameters.
  • The algorithm retrieves pixel-by-pixel optically equivalent snow grain size in two layers, snow impurity concentration, and broadband blue- and black-sky albedo.
  • The research team applied the algorithm to SGLI images obtained over the Greenland Ice Sheet, revealing a significant change in snow parameters from July 2018 to July 2019.
  • The inferred blue-sky albedo values agreed with field measurements, with a Root Mean Square Error (RMSE) of 0.0517 and a Mean Absolute Percentage Error (MAPE) of 4.64%.
  • The algorithm can easily be adapted for application to other similar multi-spectral sensors, such as MODIS, VIIRS, and OLCI.
  • The study's findings have significant implications for global climate modeling, as accurate snow parameter retrievals are crucial for understanding climate change.

Statistics:

  • The RMSE for shortwave albedo at the SIGMA-A site was 0.0517, while the MAPE was 4.64%.
  • The aligned black-sky albedo values were within 5% relative difference compared to the blue-sky values.
  • The study used SGLI images obtained over the Greenland Ice Sheet, with a significant change in snow parameters observed from July 2018 to July 2019.
  • The algorithm was validated over the Greenland Ice Sheet, with a total of 14,387 images analyzed.

Sources:

  • Snow parameter retrieval (SPR) algorithm for the GCOM-C/SGLI sensor: validation over the Greenland ice sheet. Frontiers in Environmental Science, 2025,13. (Frontiers in Environmental Science - http://www.frontiersin.org/environmental_science).
  • NewsRx. Stevens Institute of Technology Researchers Describe New Findings in Environmental Science [Snow parameter retrieval (SPR) algorithm for the GCOM-C/SGLI sensor: validation over the Greenland ice sheet]. Ecology, Environment & Conservation. May 30, 2025; p 1011.

Read more