What question did this study set out to answer?

The aim is to retrieve snow density using optical satellite data and deep learning models, advancing current methods.

April 18, 2026Remote Sensing1 citationsOpen Access

Snow Density Retrieval Based on Sentinel-2 Multispectral Data and Deep Learning

SYShuhu YangShanghai Ocean University HCHao ChenHunan University of Science and Technology YZYun Zhang

Key Points

The aim is to retrieve snow density using optical satellite data and deep learning models, advancing current methods.
Utilized Sentinel-2 multispectral reflectance data and auxiliary CanSWE measurements.
Developed a hybrid model (Snow_ACMix) integrating multi-head attention and convolutional neural networks.
Conducted feature and module ablation experiments to assess model performance.
Applied one-shot transfer learning for snow density retrieval in regions with limited data.
Achieved an MAE of 0.034 g/cm3 and R2 of 0.547 for Canada.
For Alaska, an MAE of 0.020 g/cm3 and R2 of 0.803 were determined.
The model demonstrated excellent spatial generalization with a prediction MAE of 0.027 g/cm3 in transfer learning.
Effective retrieval performance noted in mountainous regions, though influenced by wet snow conditions.

Abstract

Snow density plays a crucial role in water resource estimation, runoff forecasting, and early warning of natural disasters such as avalanches and blizzards. This study uses optical satellite multispectral reflectance data to retrieve snow density, providing a novel perspective for snow density retrieval research. Supported by auxiliary data including CanSWE in situ measurements, Sentinel-2 satellite data, and ERA5-Land reanalysis data, this study constructs a hybrid model (SnowACMix) that integrates the strengths of the multi-head attention mechanism and convolutional neural networks, realizing direct snow density retrieval from multispectral satellite reflectance data for the first time. This research was primarily conducted in Canada and Alaska. For the Canadian region, the model achieves a mean absolute error (MAE) of 0. 034 g/cm3, a root mean square error (RMSE) of 0. 051 g/cm3, and a coefficient of determination (R2) of 0. 547. For the Alaska region, the model yields an MAE of 0. 020 g/cm3, an RMSE of 0. 029 g/cm3, and an R2 of 0. 803. Feature and module ablation experiments are carried out, and one-shot transfer learning is adopted to perform snow density retrieval in the Alaska region. The spatial transfer prediction results show an MAE of 0. 027 g/cm3, an RMSE of 0. 038 g/cm3, and an R2 of 0. 747, which verify the model’s excellent spatial generalization ability and superior performance in data-scarce regions. The advantages and limitations of the SnowACMix model are investigated through comparative validation across different land cover types, regions, time periods, and against ERA5 data. The SnowACMix model achieves favorable retrieval performance in mountainous areas, and its practical application capability is verified by snow density retrieval in the Silver Star Mountain region. However, the model still has limitations: it is vulnerable to the effects of wet snow, resulting in large fluctuations in retrieval results in wet snow regions.

AIに質問

Bookmark

View Full Paper