A constrained neural differential equation trained on multi-year observations predicts snow albedo with median error under 7.5% and 10-30% improvement over prior models while generalizing to unseen sites.
A Radiative Transfer Model-Driven Machine Learning Approach to Estimate the Snow Surface Albedo Over the Greenland Ice Sheet , volume=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.app-ph 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Next-Generation Snow Albedo Parameterization for Climate Modeling using Constrained Machine Learning
A constrained neural differential equation trained on multi-year observations predicts snow albedo with median error under 7.5% and 10-30% improvement over prior models while generalizing to unseen sites.