Estimating Snow Line Altitude using Optical and SAR Data Fusion: Explainable Neural Network-Based Approach — Case Study of the Great Aletsch Glacier

Abstract

Accurate glacier surface classification is crucial for understanding glacier health. In this study, we combine Sentinel-2 optical and Sentinel-1 synthetic aperture radar (SAR) data, using an explainable neural network to determine the Snow Line Altitude (SLA). This study focuses on the Aletsch Glacier in the European Alps, which, apart from facing climaterelated retreat issues, is also affected by the presence of dust deposited during Sahara dust events. The proposed approach distinguishes pure snow from ice, aids in SLA monitoring, and also assesses the annual presence of Sahara dust on the glacier. In this paper, we observe the glacier for 2015, 2021 and 2023 and observe retreat of the SLA. The fusion of optical and SAR data mitigates the limitations of single-source data, providing a comprehensive understanding of glacier dynamics in the context of climate change.