Abstract ID: 3.10750 | Accepted as Talk | Talk/Oral | TBA | TBA

Over the last twenty years, various hazardous geomorphic processes occurred in high mountain environments, exacerbated by climate change and cryosphere degradation. These processes primarily take place during warmer months, when convectional rainfall or melt-related processes occur, and unfrozen sediments provide debris to be mobilized. Mapping their environmental impacts is important both for identifying affected areas, and for a deeper understanding of the underlying natural processes. We propose a land/water combined change detection analysis, based on satellite imagery and freely available data and tools, to be applied on occasion of extreme precipitations. Our approach is grounded in the idea that heavy rainfalls can cause two processes in mountainous areas: mobilization of debris along the hydrographic network, and sediment supply to water flows and water bodies. The first process is detected using a Land Cover Change Detection (LCCD) based on the Normalized Difference Vegetation Index (NDVI), while the second is identified through a Water Colour Change Detection (WCCD) based on water chromaticity analysis. Both LCCD and WCCD are performed separately by analysing pre- and post-event Sentinel-2 images and calculating the difference between post-event and pre-event conditions. The resulting LCCD and WCCD maps are then aggregated at catchment level counting the number of pixels that underwent land cover changes (LCCD), and calculating the average water colour difference of the water bodies within the catchment area (WCCD). Each catchment is assigned an increasing value according to its statistical deviation from the median behaviour within the area affected by the extreme event. Combining LCCD and WCCD values, a severity map of the event is generated, highlighting the catchments most affected by the rainfall’s consequences. The proposed methodology was applied to two extreme meteorological events occurred in the northern part of Italy on June 29-30 and September 4-5, 2024. It shows promise for providing a quick assessment of the impacts of extreme rainfall events across large areas (e.g. entire or multiple Italian regions). It can potentially be applied to any mountainous region worldwide, with the following main requirements: a detailed hydrographic network and Sentinel-2 imagery as cloud free as possible. Replication efforts are needed to validate the methodology’s broader applicability.