Abstract
Climate warming is driving significant changes in high-altitude mountainous regions, including rapid glacial retreat and permafrost degradation, thus leading to an increase in landslides as a paraglacial response. Identifying fresh debris deposits on the glacial surface is an important but challenging task since debris accumulation can alter glacier energy balance and interfere with alpine routes. To address this task, we propose a methodology which combines spaceborne radar and multispectral data to detect and classify glacial changes while facing the limitations of individual sensor types. The approach involves two main steps. First, cloud-free S2 images from a user-defined time range are analyzed using thresholds on the normalized difference snow index and band 8A to distinguish snow/ice from bare rock, with comparisons to the nearest maximum ablation period for improved accuracy. Second, for unclassified areas, differences in S1 backscatter intensity (Δγ0) between accumulation and ablation periods are used to detect surface changes. A size-dependent filter identifies landslide clusters by comparing Δγ0 with the previous year’s values. The final debris maps were manually validated against known landslide extents. An end to end accuracy of 90% was achieved with combined S1 and S2 data, and 70% using S1 alone. While performance decreased in highly heterogeneous or cloudcovered regions, the integrated approach improved optical-only methods. This workflow offers a robust, satellite-based solution for detecting supraglacial debris and monitoring slope instabilities, supporting climate change impact assessments in mountainous environments.